essays on the u.s. dual-class share structure jason w...
TRANSCRIPT
Essays on the U.S. Dual-Class Share Structure
by
Jason W. Howell
(Under the direction of Jeffry M. Netter)
Abstract
In the first essay, I provide a broad overview of the dual-class stock structure in the United
States. Also, I introduce the largest sample of United States dual-class firms, consisting of
1,096 firms and 8,245 firm years over the 20 year period 1988-2007. Rather than simply a
literature review, I examine the history of the structure from the early 1920s to its prevalence
in 2007. Also, I examine the implementation methods and capitalization structures used by
dual-class firms. I examine and compare certain characteristics of dual-class firms to single
class firms and briefly discuss the means firms use to exit the structure. Lastly, I survey other
control methods, such as pyramid structures and cross-ownership, and I review the financial
economic theories surrounding the dual-class structure. In the second essay, I examine a
sample 61 American dual-class firms who unify their share structure. I use the sample to
distinguish between the value recovery and optimal structure hypotheses. In line with both
hypotheses, I find a positive and significant market reaction to the elimination of the dual
class structure. In support of the optimal structure hypotheses, I find unifying firms are
inherently different than those who remain dual class. Using a probit analysis, I find unifying
firms are more likely to have lower control wedges, higher leverage and capital expenditures,
and higher levels of illiquidity. As further evidence against the value recovery hypothesis, I
find no significant change in firm value and conflicting operating performance results after
the elimination of the structure. Also, I find unifying firms are no more likely to be acquired
or taken private than their dual class counterparts. In addition, I add to the literature by
demonstrating a significant increase in liquidity for American firms who leave the dual class
structure.
Index words: Corporate Governance, One-Share/One-Vote, Dual-Class, Unifications,Private Benefits of Control
Essays on the U.S. Dual-Class Share Structure
by
Jason W. Howell
B.S., Florida Southern College, 1998
M.B.A., The University of Georgia, 2005
A Dissertation Submitted to the Graduate Faculty
of The University of Georgia in Partial Fulfillment
of the
Requirements for the Degree
Doctor of Philosophy
Athens, Georgia
2009
c© 2009
Jason W. Howell
All Rights Reserved
Essays on the U.S. Dual-Class Share Structure
by
Jason W. Howell
Approved:
Major Professor: Jeffry M. Netter
Committee: Harold Mulherin
Annette B. Poulsen
Electronic Version Approved:
Maureen Grasso
Dean of the Graduate School
The University of Georgia
2009
Table of Contents
Page
List of Figures . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . vi
List of Tables . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . vii
Chapter
1 The U.S. Dual-Class Share Structure . . . . . . . . . . . . . . . . 1
1.1 Introduction . . . . . . . . . . . . . . . . . . . . . . . . . . . . 1
1.2 History . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 2
1.3 Implementation Methods . . . . . . . . . . . . . . . . . . . . . 8
1.4 Capitalization Structures . . . . . . . . . . . . . . . . . . . . 11
1.5 Dual-Class Firm Characteristics . . . . . . . . . . . . . . . 13
1.6 Exiting the Structure . . . . . . . . . . . . . . . . . . . . . . 15
1.7 Alternative Control Mechanisms . . . . . . . . . . . . . . . 17
1.8 Financial Economic Theory . . . . . . . . . . . . . . . . . . . 18
1.9 Conclusion . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 22
2 No More Share Classes:
A Study of U.S. Dual-Class Stock Unifications . . . . . . . . . . . 28
2.1 Introduction . . . . . . . . . . . . . . . . . . . . . . . . . . . . 28
2.2 Unifications: Extant Literature and Hypotheses . . . . . . 31
2.3 Data and Sample Description . . . . . . . . . . . . . . . . . . 35
2.4 Results . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 37
2.5 Conclusion . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 46
Bibliography . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 58
iv
v
Appendix
A SEC Rule 19c-4 . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 64
B Section 313.00 of the NYSE Listed Company Manual . . . . . . . 68
C Sample of Dual Class Firms . . . . . . . . . . . . . . . . . . . . . . . 72
D Sample Unification . . . . . . . . . . . . . . . . . . . . . . . . . . . . 74
List of Figures
1.1 Dual Class Firms and % Market Capitalization by Year (1988-2007) . . . . . 23
vi
List of Tables
1.1 Number of Dual Class Firms and % Market Capitalization by Year . . . . . 24
1.2 Percentage of Dual Class Firms by Exchange . . . . . . . . . . . . . . . . . . 25
1.3 Dual Class Firms by Industry and Year . . . . . . . . . . . . . . . . . . . . . 26
1.4 Medians for Single and Dual Class Firms by Year . . . . . . . . . . . . . . . 27
2.1 Dual Class Unification Studies . . . . . . . . . . . . . . . . . . . . . . . . . . 47
2.2 Distribution of unification announcements by year and exchange . . . . . . . 48
2.3 Number of unifications by two-digit SIC code . . . . . . . . . . . . . . . . . 49
2.4 Unification Statistics . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 50
2.5 Summary Statistics . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 51
2.6 Universe of Potential Dual Class Firms . . . . . . . . . . . . . . . . . . . . . 52
2.7 Unification Event Study Results . . . . . . . . . . . . . . . . . . . . . . . . . 53
2.8 Liquidity Around Unification . . . . . . . . . . . . . . . . . . . . . . . . . . 54
2.9 Unification Determinants . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 55
2.10 Post-Unification Analysis . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 56
2.11 Firm events post-unification . . . . . . . . . . . . . . . . . . . . . . . . . . . 57
vii
Chapter 1
The U.S. Dual-Class Share Structure
1.1 Introduction
In a typical public corporation, all shareholders are provided identical voting and cash flow
rights. For example, each holder of Microsoft Corporation’s stock is allowed one vote for each
share she owns. In addition, each shareholder has residual cash flow rights to the firm and can
receive dividends. As such, shareholders are separated by the number of shares each owns;
however, the proportion of voting and cash flow rights is always proportional to the amount
invested in the firm. A shareholder who buys 10,000 shares of Microsoft stock has invested
10 times more capital in the firm than the individual who purchases only 1,000 shares. Since
each share has identical voting and cash flow rights, the holder of 10,000 shares also has 10
times more voting power.
This is not the case in a firm with two classes of stock. In a dual-class stock firm, the
investor who purchases 10,000 shares may have the same voting rights as the holder of only
1,000 shares. Voting and cash flow rights can be different based on the class of shares held.
For example, Google has two classes of stock. Class A shareholders are eligible to vote in all
corporate matters; however, they only have one vote per share, whereas class B shareholders
have ten votes per share. This allows the holders of class B shares to have control of the firm
while holding a much smaller cash flow stake. In the case of Google, co-founders Sergey Brin
and Larry Page hold zero Class A stock and 77.3% of class B shares. So although they only
own an 18.3% cash flow stake, their class B holdings give them control of the firm with a
58.3% voting stake.1
1Google, Inc., March 24, 2009 Form DEF 14A, via Edgar.
1
2
In this paper, I provide a broad overview of the dual-class structure in the United States
and introduce the largest sample of United States dual-class firms collected to date. In section
1.2, I examine the history of the dual-class structure, from the first implementation of non-
voting stock in 1898 to its use in 2007. In section 1.3, I examine the various methods firms
use to implement the dual-class structure, including methods used prior to SEC rule 19C-4.
Section 1.4 looks at the variety and frequency of the myriad voting and dividend policies used
under a dual-class capitalization. Section 1.5 looks at the general characteristics of firms who
use the structure. Section 1.6 looks at means firms use to exit the structure and section 1.7
looks at alternative methods firms use to maintain control. Section 1.8 looks at the financial
economic theory underlying the structure and section 1.9 concludes.
1.2 History
The unbundling of cash flow and voting rights dates back to the turn of the twentieth century.
Up until then, issues of both common and preferred stock were given full voting rights. It
was not until 1898 when the International Silver Company authorized twenty million shares
that non-voting stock was first issued. The authorization was for nine million preferred and
eleven million non-voting common shares. Later in 1902, the common stock was given the
right to vote; however, it was given only one vote for every two shares owned (Stevens 1926).
The non-voting stock issued by International Silver Company opened the door for firms to
begin unbundling cash flow and voting rights between common and preferred stock.
In the 1920s, firms began to issue two classes of common stock giving only one class the
right to vote. As an example, in 1925 Dodge Brothers issued 1.5 million shares of class A
non-voting stock, while the control of the firm was held by the investment bank of Dillon,
Read, and Company who owned 250,001 shares of class B voting stock. The public’s purchase
of the class A stock, bonds, and preferred stock totalled $130 million while the investment
3
bank’s controlling investment was a mere $2.25 million (Seligman 1986). By the year 1926,
at least 183 other firms had issued both class A and class B stock (Dewing 1953).2
Stock issues such as those by Dodge Brothers, Industrial Rayon Corporation, A&W Root
Beer, and Fox Theaters led Harvard University Professor William Ripley to speak publicly
about “the years of the Split Common Stock and Vanishing Stockholder.”3 His initial address
in October 1925 to the Academy of Political Science in New York City led to articles in the
New York Times, Nation, Atlantic Monthly, and to his book, Main Street and Wall Street,
published in 1927. Ripley’s railings against these “management shares” are summarized in
the following quote: “Yet the plan [dual class system] bears every appearance of a bald and
outrageous theft of the last title of responsibility for management of the actual owners by
those who are setting up these latest financial erections. Isn’t it the prettiest case ever known
of having a cake and eating it too?” (Ripley 1927).
Along with Mr. Ripley’s public speaking, scholarly articles were written by Adolf Berle
(Berle Jr 1926) and W.H.S. Stevens (Stevens 1926) addressing the one-share, one-vote con-
troversy. Mr. Ripley’s condemnation of the structure received widespread attention and the
public outcry let to the first disapproval by the New York Stock Exchange (NYSE) to an issue
of non-voting common stock on January 18, 1926. After the disapproval the NYSE issued
the following statement: “Without at this time attempting to formulate a definite policy...the
Committee...will give careful thought to the matter of voting control.” The outcry also led
President Calvin Coolidge to invite Ripley to personally discuss the issue. The February 17,
1926 New York Times headline read “President studies non-voting stocks: He confers with
professor Ripley to learn if federal action is advisable.”
After the first disapproval and statement in 1926, the NYSE prohibited the issuance of
non-voting securities, although they did not formally announce the prohibition until 1940.
2I need to look into this further because the data shows 288 firms who issued two classes from1927-1932 when the NYSE prohibition was in place. I’m having trouble locating the papers Dewingcites.
3Ripley, “From Main Street to Wall Street,” 87 Atlantic Monthly 94 (1926).
4
Between 1926 and 1985, the NYSE stock exchange kept to its prohibition with a few excep-
tions like Ford Motor Company. Ford Motor Company was able to get around the prohibition
by issuing a class with inferior voting rights rather than no voting rights. The firm’s class
B stock, which was held by the Ford family kept 40% voting power, while the class A stock
was given the remaining 60%. This allowed the family to go public while retaining control
with only 5.1% equity. Similar proportional voting structures were used in other firms such
as J.M. Smucker and American Family. Due to the strict adherence to their policy, Seligman
(1988) found only 10 NYSE firms with dual-class share structures in 1985.
Other exchanges were not as strict with their voting policies. The American Stock
Exchange (AMEX) did not implement a non-voting prohibition until 1972. In 1976, Wang
Laboratories was unable to list on the New York Stock Exchange due to its proposed dual-
class capitalization; however, the American Stock Exchange reviewed the application and
allowed the listing. This led to the AMEX issuing a policy statement on dual-class issues
(disproportionate voting rights). The key points of the statement were: 1) the limited voting
class must have the ability to elect at least 25% of the board, 2) the voting ratio should not
be greater than 10 to 1 in favor of the superior voting class, 3) no additional stock could be
issued which diluted the limited voting shareholders stake, 4) superior voting rights would
be lost if the number of shares fell below a certain percentage, and 5) dividend preference
was strongly recommended for limited voting stock. The policy became known as the “Wang
formula.”Due to their relaxed policies on the dual-class structure, Seligman (1988) estimated
approximately 7% (60 of 785) AMEX firms were dual-class in 1985, up from 37 in 1976.4
During the 1980s, the dual-class structure became a primary mechanism to prevent hostile
takeover bids. Since most firms had only a single class of stock, they implemented the dual-
class structure through various recapitalization techniques. As an example, General Cinema
Corporation performed a dual-class recapitalization by offering to exchange each common
share for a new class B share with ten votes each. The new class B share was not publicly
4Seligman (1988) also found 110 of 4101 NASDAQ companies were dual-class in 1985.
5
traded and received lower dividends than the common stock. In addition, the class B shares
were convertible to common shares but could only be transferred or sold among family
entities. As another condition, the class B shares only received 10 votes each if more than
15% of the company’s common stock is held by shareholders working in concert and if
anyone other than board members were to nominate directors.5 While the recapitalization
required shareholder approval, the company president’s family owned approximately 29% of
the common shares and the measure passed.6 With the structure setup in this manner, the
minority shareholders found it in their best interest to remain in the common share so they
could receive the higher dividend and maintain liquidity. This allowed the family to use the
new capitalization as an effective anti-takeover device.
In order to remain competitive with the American Stock Exchange (AMEX) and the
National Association of Securities Dealers (NASD) (who had no such restriction), an NYSE
subcommittee submitted a proposal in January 1985 to relax their voting policies and allow
securities with disparate voting rights to be listed as long as they met certain conditions.
Under pressure from Congress, all three exchanges then worked on a uniform policy. After
these negotiations broke down, the NYSE issued a new standard “requiring a company
proposing to recapitalize to obtain approval of the plan by a majority of its publicly held
shares, as well as a majority of its independent directors.”7
With all three exchanges now permitting dual-class structures, the structure’s use
increased and in 1988 336 firms, or 6.7% of publicly listed used the structure (see Table 1.1).
With the relaxed policies and increased use, a new call came from Congress for regulation
against the implementation of the structures. In a letter to the SEC chairman, Representa-
tive John D. Dingell, the chairman of the House Energy and Commerce Committee, stated
the “commission has the authority to mandate a one-share, one-vote rule” and that “it is
5“General Cinema board seeks new stock class to discourage suitors”, Wall Street Journal,November 14, 1984.
6“General Cinema Corp. begins exchange offer for new class B stock”, Wall Street Journal,January 2, 1985.
7“Big board ends equal vote rule”, New York Times, July 4, 1986.
6
time to move forward with sound and appropriate safeguards.”8 In his law review article on
dual-class structure, Seligman (1986) states “disproportionate voting stock is the corporate
law equivalent to price-fixing” and that “the SEC or Congress should proscribe dual class
capitalizations for the largest business corporations.”
As the discussion moved forward another view emerged. Rather than prohibit dual-class
structures all together, the focus became dual-class recapitalizations where existing share-
holders are effectively coerced into giving up their voting rights. In a New York Times
article, Steven Greenhouse asks“If management controls 55% percent of the stock and pushes
through unequal voting, is that fair to other stockholders?”9 Following this reasoning, Gilson
(1987) examines the dual-class structure and leveraged buyouts as substitutes and concludes
“a resolution-prohibition of dual class transactions but not dual class capital structures-
becomes apparent. That resolution would leave intact the benefits of the dual class capital
structure, while still preventing any dominant shareholder group from using dual class trans-
actions [recapitalizations] to coerce a firm’s public shareholders.” This new approach led to
the proposal of rule 19C-4 by the SEC.
On July 7, 1988, the Securities and Exchange Commission voted 4 to 1 to implement rule
19C-4 (see Appendix A for the full text of rule 19C-4). Under the rule, the SEC prohibited
self-regulatory organizations from listing and trading the stocks of any company that issued
new shares carrying more than one vote per share, but it allowed companies to issue shares
with less than one vote per share and permitted those with unequal voting rights to still be
traded. As soon as the new rule was passed, questions were raised as to whether the SEC had
the legal authority to enforce such policies on self regulating organizations such as the NYSE
and NASD. On June 12, 1990, a three judge panel of the United States Court of Appeals for
the District of Columbia Circuit unanimously ruled the SEC had exceeded its authority.
Despite the court’s rejection of 19C-4, the NASD proceeded with implementing a 19C-4
type rule allowing firms to introduce inferior voting shares during initial public offerings
8“Unequal stock class opposed”, New York Times, May 24, 1988.9“Unequal votings rights in stock”, New York Times, March 19, 1985.
7
but barring firms from reducing existing shareholders’ voting rights.10 The NASD joined the
NYSE who had already voluntarily issued a policy implementing the 19C-4 rule. In June
1991, the AMEX moved to restrict its policy similar to the 19C-4 rule, with the exception
that inferior voting shares could be created if approved by two-thirds of the stockholders and
a majority of non-insiders.11
In December 1993, SEC Chairman Arthur Levitt Jr. suggested all U.S. markets implement
a uniform policy regarding voting rights.12 In line with Mr. Levitt’s suggestion, the AMEX
and NASD shortly thereafter approved a uniform policy which was followed by the NYSE in
May of 1994. The voting policy (see Appendix B for full text) allows companies to be listed
who have dual classes of stock and sets no restrictions on voting rights for new public offerings
of stock. However, it bars companies from taking steps to reduce their existing shareholders’
voting rights through such actions as“the adoption of time phased voting plans, the adoption
of capped voting rights plans, the issuance of super voting stock, or the issuance of stock
with voting rights less than the per share voting rights of the existing common stock through
an exchange offer.”
Despite the exchanges developing a uniform policy and preventing coercion through dual-
class recapitalizations, the structure still receives criticism. The critics call for a one-share, one
vote standard and point to the anti-takeover property of the structure, the risk of entrench-
ment, and potential expropriation of minority shareholders. They contend insiders with con-
trol will take on bad projects, reject sound takeover offers, or just not run the firm effectively.
In 2004 when Google went public with the dual-class structure , Bob Monks, shareholder
activist, stated “It is stupid to have two classes of stock. I think they have been badly
advised.”13 Charles Elson, director of the John L. Weinberg Center for corporate governance
10“NASD plans a one-share, one-vote rule”, Wall Street Journal, June 21, 1990.11“AMEX files plan for holders’ votes on classes of stock”, emphWall Street Journal, June 13,
1991.12“NYSE approves shareholder voting rights policy”, Dow Jones News Service, May 5, 1994.13Foremski, London, and Waters, “Google and the establishment set to clash”, Financial Times,
May 1, 2004.
8
at the University of Delaware, added “I think it is a terrible mistake. Any time you separate
ownership from control there is trouble down the line.”14
In recent years, institutions, unions, and blockholders have led shareholder proposals
to eliminate the structure and move to one vote per share. For example, in 2007 John
Chevedden led a proposal to remove the dual-class structure at Ford Motor Company. In
the proposal he states “Dual-class stock companies like Ford take shareholder money but do
not let shareholders have an equal voice in their company’s management. Without a voice,
shareholders cannot hold management accountable. Shareholders who finance our company
should be able to hold our management accountable.”15 The initiative was opposed by the
board, yet garnered support from 27% of shareholders.16 Similar proposals have been rejected
recently at firms such as the New York Times, Google, Emmis, and Sotheby’s.
Although the structure has received criticism since the 1920s and the risk of hostile
takeovers have dramatically declined since the 1980s, approximately 7.4% of publicly traded
firms still choose to use the dual-class structure, including firms such as Google, News Cor-
poration, and Blockbuster (see Table 1.1).
1.3 Implementation Methods
There are a number of methods firms can use to implement the dual-class share structure.
Before the SEC introduced rule 19C-4, companies commonly introduced the dual-class struc-
ture through a recapitalization. A common method firms used to recapitalize was through
the use of a“dividend sweetener.”With this method, a firm with one class, creates a new class
with less voting power but with higher dividends. The firm then gives existing shareholders
the option to convert to the new inferior voting class with higher dividends. For minority
shareholders, who do not have enough shares to affect decisions, it is in their best interests
14Foremski, London, and Waters, “Google and the establishment set to clash”, Financial Times,May 1, 2004.
15Ford Motor Company, April 5, 2007 Form DEF 14A, via Edgar.16Stoll, “Ford shareholders take swipe at family voting power”, Dow Jones Newswires, May 10,
2007.
9
to move to the newly formed restrictive voting class to take advantage of the dividend pref-
erence and leaves blockholders holding the superior voting shares, allowing them to maintain
control of the firm. For example in 1984, BDM International offered its shareholders 1.1
shares of new class A stock for each existing share. The class A stock would receive 15%
higher dividends but would only receive 0.10 votes per share and could only elect 25% of the
board members. The shares that were not swapped were converted to class B shares which
had full voting rights and could vote on 75% of the board. In addition, the firm announced
prior to the conversion that the class B shares would be delisted after the conversion. This
added more pressure to minority shareholders to convert to class A.17
The “dividend sweetener” method was also used in combination with a two for one stock
swap. In 1988, Concord Fabrics’ shareholders ratified a plan to issue a class A share and a
class B share for each share owned. The class A shares were promised a higher dividend rate
and liquidation value, but were only given one vote per share. Class B shares were given
ten votes per share. The Weinstein family owned 62% of the stock so it was in their best
interests to hold the class B shares, while it was in minority shareholders best interests to sell
their class B shares for class A shares.18 This structure allowed Concord Fabrics to effectively
segregate voting power from minority shareholders.
Another method specifically restricted in rule 19C-4 was the use of time phased voting
plans. With this method, shareholders were segregated and voting rights distributed based
on the length of time the shareholders owned the stock. In 1985, the shareholders of J.M.
Smucker Company passed a proposal that gave shareholders 10 votes per share on the con-
dition they had held the share continuously for at least four years. Using this method, the
company attempted to get around the NYSE’s dual-class restriction because it did not create
two classes of stock but two classes of holders.19
17“BDM extends offer to swap new stock issue for common”, Wall Street Journal, January 24,1984.
18“Shareholders ratify measure creating 2 classes of stock”, Wall Street Journal, April 5, 1988.19“J.M. Smucker Co. holders consider anti-takeover step”, Wall Street Journal, August 1, 1985.
10
Since SEC rule 19C-4 and the changes in SRO rules, most firms choose to implement the
dual-class structure during their initial public offering. In these cases, the management of the
firm desires to retain control while also accessing the capital markets to fund positive NPV
projects. Smart and Zutter (2008) find 9.6% of 2,622 initial public offerings are done with
dual classes of stock during 1990 to 1998. Recent examples of firms who have implemented
the structure at their IPO are Google and Rosetta Stone Software.
In some cases the dual class structure is implemented during a spinoff to minimize a tax
burden. Before 1998, a Morris Trust enabled a firm to receive favorable tax treatment if
they retained 80% of the voting rights of the new firm in a spinoff. For example, in 1997
Hughes Electric was spun off from General Motors and merged with Raytheon. In order to
obtain special tax treatment, General Motors had to retain 80% of the voting power in the
new firm. This was accomplished by the new firm issuing two classes of stock that allowed
General Motors to retain 80% of the voting rights.20
Firms also move to the dual class structure by issuing a stock dividend. For example, on
June 12, 1994 Cherry Corporation converted its existing common shares to class B voting
shares and four days later the board of directors authorized a stock dividend of one class A
non-voting share for each class B share owned. The stock dividend occurred on July 11, 1994
and the next day the firm filed form S-2 to issue additional class A non-voting shares. The
new issue concluded on August 19, 1994 and the firm received $33 million in net proceeds.21
By using this method of implementation, the original shareholders maintained control of the
firm while at the same time accessing new capital for the firm. Other examples of firms who
moved to the dual class structure using this method are Dow Jones & Co, Times Mirror
Company, CMI Corporation, and Baker Corporation.
20“For some companies, A+B=1.” CFO.com. February 13, 2001.21Cherry Corp., February 28, 1995 Form 10-k (filed May 22, 1995), via Edgar.
11
1.4 Capitalization Structures
While the most common dual-class structure is a 10:1 voting ratio with two classes of stock,
firms unbundle voting and cash flow rights in many different ways. In this section, I discuss the
various voting, dividend, and convertibility clauses firms use within the dual-class umbrella.
Firms who unbundle cash flow and voting rights are typically referred to as dual-class
firms, although it does not necessarily mean the firm has only two classes of stock. Gompers,
Ishii, and Metrick (2008), find 28 firms with more than two classes of stock in their eight
year sample from 1995 to 2002, this includes at least six firms with four classes of stock. For
example Comcast Corporation has three classes of stock: class A with one vote per share,
class B with 15 votes per share, and class A special with no voting rights.22 Radio One and
Univision Communications are examples of firms with four classes of stock.
Gompers, Ishii, and Metrick (2008) find the most common voting structure setup is the
10:1 voting ratio.23 Approximately 36% of dual class firms use the 10:1 voting ratio, where
the superior voting class receives 10 votes per share and the inferior voting class receives
only 1 vote per share. Examples include Google, American Greetings, and Dow Jones. 15%
of dual class firms use a voting ratio higher than 10:1. For example, the class A shares of
Boca Resorts, a firm controlled by Wayne Heizunga, are entitled to one vote each, while
the class B shares are entitled to 10,000 votes each.24 A more modest example is the Coca-
Cola Bottling Company where the class B shares are entitled to 20 votes each.25 For 18%
of dual-class firms, the voting ratio is less than 10:1. For example, Blockbuster’s class B
stock is entitled to two votes per share, while the class A stock is entitled to one.26 The
remaining 31% of dual-class firms, use proportional voting for directors. In these cases, both
classes have one vote per share but the inferior voting class can only elect a minority of the
22Comcast Corp., December 31, 2008 Form 10-k (filed February 20, 2009), via Edgar.23I use the 2002 data for percentages.24Boca Resorts, June 30, 2004 Form 10-K, (filed September 13, 2004), via Edgar.25Coca-Cola Bottling Company, December 28, 2008 Form 10-K, (filed March 13, 2009), via Edgar.26Blockbuster, Inc., January 6, 2008 Form 10-K, (filed March 6, 2008), via Edgar.
12
directors (typically 25%). An example is the Washington Post. Their minority voting class,
class A, is entitled to only vote on 30% of the board of directors.27
Since dual-class firms are unbundling both cash flow and voting rights, not only do they
use various voting structures but they may also have disproportionate dividend policies. In
the 1980’s, the minority voting class was commonly given higher dividends as a conversion
sweetener; however, since most firms now become dual class at the initial public offering
this is no longer the case. Gompers, Ishii, and Metrick (2008) finds only 13% of dual-class
firms give the inferior voting class a higher dividend. The Hershey Company is an example
of one of the 13% whose inferior voting class is entitled to a higher dividend. Since their
dual-class structure implementation in 1984, the Hershey Company has given inferior voting
shareholders a 10% higher dividend than class B superior voting shareholders.28 According
to Gompers, Ishii, and Metrick (2008), approximately 86% of dual-class firms pay equal
dividends to superior and inferior voting classes. A very small minority of 1% (five firms)
actually give the superior class a higher dividend than the inferior class.
Each dual-class firm has at least two classes of stock authorized and issued; however, gen-
erally only the restricted voting class is traded publicly. Gompers, Ishii, and Metrick (2008)
find that both classes are traded publicly only 15% of the time. Take Google and the New
York Times for example, only the restricted voting class is traded publicly; however, Berk-
shire Hathaway trades both class A and B publicly. Since 85% of dual-class firms trade only
one class publicly, oftentimes the bylaws allow superior vote holders to convert their superior
voting shares one-to-one for inferior voting shares. This allows superior voting shareholders
to maintain liquidity. In addition, many firms setup automatic conversion features if a certain
threshold of superior voting shares are converted. For example, if class B holdings fall below
10% they will automatically convert to class A.
27Washington Post, December 28, 2008 Form 10-K, (filed February 26, 2009), via Edgar.28Hershey Company, December 31, 2008 Form 10-K, (field February 20, 2009), via Edgar.
13
1.5 Dual-Class Firm Characteristics
Since only about 7% of firms use the dual-class structure, which firms use the dual-class
structure? Are there any unique characteristics about the firms who choose the structure? In
this section, I examine dual-class firm characteristics and where applicable how they compare
to single class firms.
In the 1970s, Seligman (1986) finds the American Stock Exchange had the least restrictive
policy towards dual-class firms and therefore had the highest percentage of dual-class firms
(approximately 7%). In addition, he found the NYSE had the least number of dual-class firms
due to its restrictive voting rights policy. However since SEC rule 19C-4 was introduced in
1988, the NYSE has increasingly picked up dual-class firms (see Table 1.2). In 2007, I find
11.5% of firms on the New York Stock Exchange have dual class structures, while only 5.1%
of AMEX and 5.5% of NASDAQ firms use the structure.
With the implementation of rule 19C-4, firms were restricted in the manner in which
they could become dual-class. Since investors knew beforehand what they were buying there
were no restrictions placed on initial public offerings. As such, an increasing number of firms
now use the dual-class structure at their IPO. Smart and Zutter (2003) examine dual class
initial public offerings from 1990 to 1998, and find an increase in the use of the structure.
They find between 1994 and 1998, 11% of all firms used the dual-class structure and these
11% made up 33% of the IPO market capitalization.
The dual-class structure separates cash flow and voting rights between stockholders. Thus,
the dual-class structure allows for insiders to hold a large percentage of voting rights, while
minimizing their cash flow stake. In the case of Google, this wedge between voting and cash
flow rights is approximately 40% (58.3% voting, 18.3% cash flow). However, according to the
evidence the wedge is not always so large. Gompers, Ishii, and Metrick (2008) find insiders
hold on average 60% of voting rights and 40% of cash-flow rights. They find that in only
one-third of dual class firms do the insiders maintain a majority of voting rights but do not
hold a majority of cash flow rights.
14
The dual-class structure allows insiders to keep control of the corporation, and based on
theory, we should find the dual-class structure in industries in which their are significant
benefits to maintaining control. According to Demsetz and Lehn (1985), the amenity poten-
tial of a firm’s output is one determinant that could lead to increased control or ownership
concentration. In line with this, I find a clustering of dual-class firms in the communications
and publishing industries (see Table 1.3). In addition, I find the business services, food and
kindred products, and chemicals industries have a high number of dual-class firms.
On average, dual-class firms are larger and more highly leveraged than single class firms.
Table 1.4 shows the medians for single and dual-class firms for the years 1988, 1992, 1996,
2000, 2004, and 2007. Measured by assets, dual-class firms are significantly larger than single
class firms and the difference in size has risen over the years. Also, dual-class firms have
significantly larger debt-to-asset ratios than single class firms. Once again, the difference has
been significant since 1988 and they hold more than twice the leverage of single class firms
in 2007. These results are in line with previous studies such as Gompers, Ishii, and Metrick
(2008). Gompers, Ishii, and Metrick (2008) examine dual class and single class firms in 2000
and finds dual class firms to have average mean assets of $3.1 billion while the single class
mean is only $2.1 billion. In addition, they find dual class firms have on average 35% higher
leverage.
In addition to size and leverage, Table 1.4 shows dual-class firms have significantly lower
research and development to sales ratios, higher return-to-assets ratios, and significantly
lower Tobin’s q. Although the analysis is simply univariate, the Tobin’s q result does corre-
spond to previous research that shows the dual-class structure has a negative effect on firm
value as measured by Tobin’s q (Gompers, Ishii, and Metrick (2008) and Claessens, Djankov,
Fan, and Lang (2002)). The significant difference in the return-to-assets ratio in 2000, 2004,
and 2007 is contrary to previous research on the accounting returns (Smart, Thirumalai,
and Zutter (2008)) who find no significant difference in accounting returns between dual and
single class firms.
15
Gompers, Ishii, and Metrick (2008) examine the determinants of dual-class status. They
find the most powerful determinant of the dual-class structure is whether a person’s name
appears in the name of the firm at the initial public offering. They also find four other
determinants: whether the firm is in a media industry, the number and size of firms in the
same metropolitan area, and the sales of the firm compared to others going public in the
same industry.
1.6 Exiting the Structure
The dual-class structure is implemented in various ways and is also eliminated in various
ways. Firms leave the structure during mergers, acquisitions, and during bankruptcy but
they also eliminate the structure in other means.
A common policy implemented with a dual-class structure is the condition that if the
voting class drops to a certain percentage of the total number of shares voting for the company
then the superior voting shares will automatically convert to inferior voting shares. At Zebra
Technologies they had a dual class structure with the standard 10:1 voting ratio, with class B
getting 10 votes. Only class A shares traded publicly and shareholders could exchange B for A
(1 for 1) at their option. The certificate of incorporation stated that if the outstanding shares
of Zebra class B stock cease to represent 10% of the total number of shares, then the Class
shares would automatically convert to class A. On July 1, 2003 the class B shares dropped
below 10% and the subsequent class B shares were automatically converted, eliminating the
dual class structure at the company.29
Some firms exit the dual class structure when the majority shareholder sales or exchanges
their superior voting shares. In 2005, the Taubman family exchanged their 14 million shares
of Sotheby’s class B (superior voting) stock for $168 million in cash and 7.1 million Class A
shares. The family did not own all of the class B shares (approximately 4 million held by
others); however, the firm had a clause that if the voting power of the class B shares fell below
29Zebra Technologies, December 31, 2003 Form 10-K, (filed February 27, 2004), via Edgar.
16
50% they would automatically be converted to class A shares. When the Taubman family
converted their shares, the threshold was passed and the dual-class structure was eliminated.
According to the company, they believe “the simplified share structure will enhance share
liquidity and increase the Company’s strategic and financing flexibility.”30
In some cases, firms exit the dual-class structure by moving to the standard single class
structure through a shareholder proposal. This is commonly referred to as a unification.
Howell (2009) finds 61 dual-class firms unified their share classes during the years 1992-2006.
In most cases, the board meets and examines the costs and benefits of moving to a single
class structure. Once approved by the board, the proposal is put to a shareholder vote. The
most common conversion rate is for both superior and inferior voting shares to receive one
share of the new common stock.
GameStop Corporation performed a unification in February 2007. GameStop’s dual-class
structure consisted of Class A shares with one vote per share and class B shares with ten
votes per share. Both share classes were traded publicly; however, the class B shares were
not convertible to class A shares.31 In the simplification of the share structure each class B
shareholder received a class A share.32 In the firm’s proxy statement, they state “the Board
has determined that it is in the best interests of the Company and its stockholders to convert
the class B common stock into class A common stock to simplify its capital structure, improve
corporate governance by reducing the appearance of being a closely-held company, reduce the
expenses and confusion associated with maintaining two separate classes of common stock
and improve liquidity and trading volume of outstanding shares.”33
30Sotheby’s Holdings Inc., December 31, 2005 Form 10-K, (filed March 16, 2006), via Edgar.31In their proxy statement, they had approximately 46 million shares of class A stock held by
61 stockholders of record, and approximately 30 million shares of class B stock held by 1,363stockholders. The class B stock made up 39.4% of the cash-flow rights and 86.7% of the votingrights.
32GameStop Corporation, February 2, 2008 Form 10-K, (filed April 2, 2008), via Edgar.33GameStop Corporation, December 29, 2006, Form DEF 14A, via Edgar.
17
1.7 Alternative Control Mechanisms
The dual class structure allows a firm’s blockholders to maintain control of the firm through
voting rights while holding less cash-flow rights than would be required to typically keep
control. This type of mechanism is referred to as a controlling minority structure (CMS).
Other controlling minority structures are pyramids and cross ownership.
With a pyramid structure, a controlling stake is held in a holding company, which in turn
holds a controlling stake in another firm. For example, a controlling stakeholder holds a 50%
plus one stake in company A and company A holds a 50% plus one stake in company B, then
the controlling stakeholder maintains control of company B with only a 25% equity stake
in the firm. This pyramid can continue for multiple levels and can reach multiple firms. For
example, La Porta, Lopez-de Silanes, and Shleifer (1999) show how the Wallenberg family,
in Sweden, controls ABB, the fourth largest firm in Sweden by market capitalization by
only holding a 5% cash flow stake. Pyramid structures are not frequently used in the United
States; however, they are commonly used in countries outside of the United States.
Another alternative to maintain control is cross-ownership. Cross-ownership occurs when
a firm owns portions of other firms in which it does business. In this way, the management
group can maintain tight control of the firm through its relationships with its other compa-
nies. In the United States, cross-ownership is not used as frequently, due to legal restrictions.
For example, there is currently a restriction on newspaper-broadcast cross-ownership, which
the FCC attempted to eliminate but Congress nullified.
One way the dual class structure allows insiders to maintain control is by its effective
anti-takeover property. With the dual class structure in place, it makes for an effective anti-
takeover mechanism by allowing the controlling group to veto any takeover proposals. Other
anti-takeover mechanisms that are used include poison pills and leveraged buyouts. A poison
pill seeks to dissuade buyers by implementing detrimental plans if taken over. For example,
a shareholder rights plan is a type of poison pill that will dilute the bidder’s ownership
percentage if a takeover occurs. The dilution occurs by automatically increasing the previous
18
shareholders shares. A leveraged buyout occurs when a management group takes a firm
private by using debt.
1.8 Financial Economic Theory
In the United States, firms typically have a single class of common stock. An owner of this
common stock holds a right to a proportion of the residual cash flows of the firm based on
his/her proportion of ownership. Since the stockholder has a right to these residual cash flows,
he/she has incentive to ensure there will in fact be residual cash flows. In other words, it is
in the stockholder’s best interest to ensure the firm satisfies obligations to stakeholders and
provides a desirable product, while at the same time minimizing cost and thus maximizing
residual cash flows. It is for this reason that the residual claimants are giving the ability to
monitor the firm through the use of voting (Alchian and Demsetz 1972).
With the bundling of cash flow and voting rights, the stockholder is given a mechanism
to affect residual cash flows. However, how and when are residual cash flows paid? They
are paid through dividends and at the liquidation of the firm. While dividends may be paid
every quarter (if they are paid at all), liquidation cash flow payments only occur when it
is necessary for the firm to liquidate. So for the residual claimant, their rights must go
beyond cash flow and voting rights, they must have the right to sell their claim (Alchian and
Demsetz 1972).
Under a single class structure, residual claimants have all three rights, the right to cash
flows, voting, and selling, and these rights are equal among all stockholders based on their
proportion of stock ownership. In a firm with a dual class structure, voting rights and cash
flow rights are segregated based on the class of share owned by the stockholder. One class
may have ten votes per share, while the other class may have one or even no votes per share.
By deviating from the standard one-share, one-vote structure, firms lose the binding link
between cash flows and monitoring, and can restrict the shareholder from receiving full value
for his or her share.
19
This separation of cash flow and voting rights leads to an exacerbation of the problems
associated with the separation of ownership and control in firms with a single class structure.
As ownership dispersion increases those who own the firm tend to be different than those who
control the firm (managers). This separation leads to conflicts between the agent (manager)
and the principal (owner) and agency costs are incurred (Berle and Means (1932), Jensen
and Meckling (1976)). Since the agent does not have a cash flow stoke, he/she has no inherent
desire to maximize residual cash flows and takes actions to secure their positions.
In a similar manner, firms with a dual class structure have a separation between those
with voting control and those with cash flow rights. This leads to conflicts of interests and
agency costs (Masulis, Wang, and Xie (2008). Those with voting control seek to retain their
control through entrenchment. At the same time, they have less interest in the cash flows of
the firm, therefore they are more likely to expropriate other shareholders (those with cash
flow rights). On the other hand, shareholders with cash flow rights seek to maximize the
value of these cash flows without regard to the control of the firm. Thus leading them many
times to be in direct conflict with voting control holders on issues such as mergers, takeovers,
and directors.
With the agency problems associated with ownership dispersion, why allow it in the
first place? During a firm’s growth cycle, there reaches some point in time where the firm’s
available capital, either through retained earnings or the firm’s owners, is not sufficient to
finance future growth. It then becomes necessary for the firm to reach out to the capital
markets, through the use of debt and external equity. In the same way, firms choose to issue
restricted voting equity so they may retain control of the firm while also raising capital to
finance future projects. Also, a firm’s owners may have sufficient capital to finance future
growth but may choose to withhold funds so they can diversify their investments.
According to DeAngelo and DeAngelo (1985), the dual class structure is an “intermediate
organizational structure which fits somewhere between the polar cases of the dispersed-
ownership public corporation and the closely held firm.” With this intermediate organiza-
20
tional structure, the dual class structure faces agency problems not only between owners and
managers, but between owners of voting and cash flow rights. This leads to trade-offs between
the benefits of control and monitoring and the costs of entrenchment and expropriation.
1.8.1 Empirical Findings
With the trade-offs facing firms using the dual-class structure, the first question is how does
the structure affect shareholders. Does the implementation of the structure lead to a decrease
in stock price due to the structure’s entrenchment and expropriation properties or do stock-
holders view the structure as just simply a mechanism the firms uses to fund growth? Partch
(1987) examines 44 U.S. recapitalizations between 1962 and 1984 and determines there is no
significant change in shareholder wealth due to the dual class structure’s implementation.
Jarrell and Poulsen (1988) extend the U.S. sample to 94 firms and finds significant negative
abnormal returns (-0.82%) for firms recapitalizing. In the most comprehensive study to date,
Dimitrov and Jain (2006) use a sample of 178 recapitalizations from 1979 to 1998 and find
a slightly positive three-day abnormal return (-0.06%) that is not significant.
The dual-class structure introduces new “agency problems” into the firm by separating
ownership between superior voting and inferior voting holders. Masulis, Wang, and Xie (2008)
finds four specific areas where the separation of voting and cash-flow rights incur costs at the
expense of inferior voting shareholders.34 Based on these additional costs, we would assume
firm performance would be worse at dual-class firms when compared to single class firms.
However, some evidence points to an increase in performance after the implementation of
the structure. Lehn, Netter, and Poulsen (1990) find positive industry-adjusted operating
performance for firms who recapitalized between 1978 and 1987. Also, Dimitrov and Jain
(2006) find holders of dual-class stock firms earn an average abnormal return of 23.11% in
the four years following recapitalization.
34These are corporate cash valued less to outside stockholders, CEOs receive higher compensation,managers are more likely to make value-destroying acquisitions, and capital expenditures contributeless to shareholder value.
21
However, in studies using Tobin’s Q as a proxy for firm value results tend to show a neg-
ative affect of firm value as the wedge between voting and cash flow rights increases. Based
on a sample of U.S. dual-class firms between 1995 and 2002, Gompers, Ishii, and Metrick
(2008) find firm value is positively associated with insiders’ cash-flow rights and negatively
related to insiders’ voting rights when using single-stage regressions. In addition, they find
firm value negatively associated with the wedge between voting and cash flow rights. When
they control for endogeneity, estimates are similar to the single-stage regressions; however,
the significance is lower. Similarly, Villalonga and Amit (2006) find control enhancing mech-
anisms, such as dual-classes, pyramids, and cross-ownership, have a negative and significant
effect on firm value for S & P 500 firms during 1994 to 2000. Consistent with this result,
Villalonga and Amit (2008) find the wedge between voting and cash-flow rights is negatively
associated with firm value for family firms; however, the result is not significant.
The empirical evidence to date shows that U.S. dual-class firms who unify their share
classes have positive announcement effects. Smart, Thirumalai, and Zutter (2008) examine
the unification announcement effects of 37 firms who chose to unify their share classes. They
find a positive and significant five-day announcement effect of 2.7%. Consistent with Smart,
Thirumalai, and Zutter (2008), Howell (2009) examines 61 unification events between 1992
and 2006 finds a positive and significant announcement effect for restricted voting and shares
(1.85%) and a total significant effect of 2.23% when including both restricted and superior
voting classes.
1.8.2 Structure Used to Examine Theories
Since the dual-class structure provides an“extreme form”of corporate governance (Gompers,
Ishii, and Metrick 2008). The structure provides a suitable testing ground for many theories
and ideas in corporate finance.
Barclay and Holderness (1989) introduce the idea that if all shareholders receive benefits
directly proportional to their ownership stake, sales of large blocks of stock should be priced
22
at the same price as smaller sales of stock. They hypothesize that large blocks would trade at a
premium if there are “private benefits” that accrue to these large blockholders. They examine
63 block trades and find an average premium of 20% which questions the assumption that
all stockholders are homogeneous and benefits are proportional to ownership. They attribute
this premium large blockholders of private benefits of control. Zingales (1994) and Zingales
(1995) uses the unique property of dual-class shares to expand the research on private benefits
of control. In essence the premium that superior voting shares receive over inferior voting
shares is made up at least partly by the private benefits of control that accrue to superior
vote holders. Based on this technique, studies such as Doidge (2004) and Nenova (2003)
examine the value of a vote and private benefits of control across countries.
Are institutional investors concerned about corporate governance within a firm or do
they simply chase past returns? Li, Ortiz-Molina, and Zhao (2008) examine institutional
investment in dual-class firms and find that institutional investment is significantly lower in
dual-class firms than in the standard single-class firm. Also, they find institutional investment
increases when dual-class firms unify their share classes. This evidence demonstrates that a
firm’s corporate governance matters affects the decisions of institutional investors.
1.9 Conclusion
The dual-class structure allows the separation of voting and cash-flow rights among owners.
It allows blockholders to maintain control while accessing the external equity market, but
can lead to expropriation of minority shareholders and entrenchment. Although the structure
has faced opposition over the years, from Mr. Ripley in the 1920s to SEC Rule 19C-4 in the
late 1980s, the structure continues to be used in 7.4% of publicly traded firms in the United
States and has proven to be a viable corporate governance structure.
23
Figure 1.1: Dual Class Firms and % Market Capitalization by Year (1988-2007)
24
Table 1.1: Number of Dual Class Firms and % Market Capitalization by YearThis table shows the number of firms by fiscal year. Dual-class firms are identified by the processdescribed in Appendix C. Market capitalization figures are based on the closing fiscal year price(PRCCF) and number of common shares outstanding (CSHO) reported by CRSP/Compustat.For dual class firms, the number of common shares outstanding is based on the sum of all classes.Data comes from the CRSP/Compustat Fundamentals Annual database.
Number of Firms Market Capitalization (Billions)Year All Dual Class Percentage All Dual Class Percentage1988 5,013 336 6.7% $ 2,427 $ 150 6.2%1989 4,911 339 6.9% $ 2,926 $ 178 6.1%1990 4,893 348 7.1% $ 2,721 $ 150 5.5%1991 5,030 359 7.1% $ 3,589 $ 189 5.3%1992 5,250 377 7.2% $ 3,990 $ 227 5.7%1993 6,256 409 6.5% $ 4,605 $ 304 6.6%1994 6,555 441 6.7% $ 4,568 $ 291 6.4%1995 6,685 455 6.8% $ 6,170 $ 382 6.2%1996 7,112 511 7.2% $ 7,517 $ 517 6.9%1997 7,079 509 7.2% $ 9,926 $ 705 7.1%1998 6,697 502 7.5% $ 11,958 $ 797 6.7%1999 6,461 504 7.8% $ 15,018 $ 1,435 9.6%2000 6,298 497 7.9% $ 14,961 $ 1,122 7.5%2001 5,760 447 7.8% $ 12,691 $ 1,154 9.1%2002 5,399 417 7.7% $ 10,037 $ 965 9.6%2003 5,096 387 7.6% $ 12,850 $ 1,190 9.3%2004 5,025 376 7.5% $ 14,330 $ 1,360 9.5%2005 4,945 362 7.3% $ 14,836 $ 1,415 9.5%2006 4,871 336 6.9% $ 16,493 $ 1,608 9.8%2007 4,530 333 7.4% $ 16,686 $ 1,566 9.4%
Total 113,866 8,245 7.2%
Firms 13,971 1,096 7.8%
25
Table 1.2: Percentage of Dual Class Firms by ExchangeThis table shows the percentage of firms in each exchange that are dual class. Dual-class firms areidentified by the process described in Appendix C. For example in 2000, 12.3% of all NYSE firmshad dual classes of stock. The firm’s exchange is identified by the exchange (EXCHG) code in theCRSP/Compustat Fundamentals Annual database.
Year NYSE AMEX NASDAQ OTC1988 6.7% 11.1% 7.5% 5.1%1989 6.9% 11.7% 7.1% 5.7%1990 7.5% 12.5% 6.7% 5.9%1991 7.9% 12.0% 7.4% 5.0%1992 8.2% 11.2% 7.3% 5.0%1993 8.0% 10.0% 6.1% 4.9%1994 8.1% 11.0% 6.1% 5.3%1995 8.4% 10.6% 6.2% 5.4%1996 10.0% 9.6% 6.2% 5.6%1997 10.3% 9.3% 6.1% 5.6%1998 11.0% 8.7% 6.1% 6.0%1999 12.1% 8.6% 6.1% 6.4%2000 12.3% 8.7% 6.3% 6.4%2001 11.9% 8.9% 6.1% 6.3%2002 11.9% 7.8% 6.0% 6.2%2003 11.5% 7.6% 5.8% 6.3%2004 11.4% 7.1% 5.7% 5.9%2005 11.4% 6.3% 5.6% 4.9%2006 11.2% 5.6% 5.2% 3.9%2007 11.5% 5.1% 5.5% 5.6%
26
Table 1.3: Dual Class Firms by Industry and YearThis table shows the number of dual class firms in each two-digit SIC code in the years 1988, 1992,1996, 2000, 2004, and 2007. Dual-class firms are identified by the process described in AppendixC. Data comes from the CRSP/Compustat Fundamentals Annual database.
SIC Description 1988 1992 1996 2000 2004 200701 Agricultural Production - Crops 2 3 2 3 202 Agricultural Production Livestock and Animal Special-
ties1 1 1 1
07 Agricultural Services 110 Metal Mining 1 3 3 2 112 Coal Mining 113 Oil and Gas Extraction 8 12 7 8 7 1314 Mining and Quarrying Of Nonmetallic Minerals, Except
Fuels1 1 1
15 Building Construction - General Contractors and Oper-ative Builders
3 5 7 5 3 3
16 Heavy Construction Other Than Building Construction- Contractors
2 1 1
17 Construction - Special Trade Contractors 3 3 3 3 2 120 Food and Kindred Products 22 19 23 21 19 1721 Tobacco Products 222 Textile Mill Products 9 10 7 7 3 323 Apparel and Other Finished Products Made From Fab-
rics and Similar Materials3 2 4 6 5 4
24 Lumber and Wood Products, Except Furniture 1 2 2 2 1 125 Furniture and Fixtures 3 4 4 5 4 326 Paper and Allied Products 3 3 5 4 2 227 Printing, Publishing and Allied Industries 20 26 24 18 17 1528 Chemicals and Allied Products 21 20 24 22 16 1529 Petroleum Refining and Related Industries 2 2 230 Rubber and Miscellaneous Plastics Products 4 7 7 7 6 231 Leather and Leather Products 2 3 5 5 5 432 Stone, Clay, Glass and Concrete Products 3 2 2 2 1 133 Primary Metal Industries 3 3 9 8 1 234 Fabricated Metal Products, Except Machinery and
Transportation Equipment4 3 8 7 7 4
35 Industrial and Commercial Machinery and ComputerEquipment
15 18 25 17 12 9
36 Electronic and Other Electrical Equipment and Compo-nents, Except Computer Equipment
25 26 28 27 19 16
37 Transportation Equipment 9 9 13 12 9 738 Measuring, Analysing, Controlling Instruments; Photo-
graphic, Medical & Optical Goods; Watches & Clocks16 16 15 10 6 6
39 Miscellaneous Manufacturing Industries 5 5 6 5 5 540 Railroad Transportation 2 3 2 141 Local and Suburban Transit and Interurban Highway
Passenger Transportation1
42 Motor Freight Transportation and Warehousing 1 5 6 6 244 Water Transportation 1 1 245 Transportation By Air 2 1 3 1 1 146 Pipelines, Except Natural Gas 2 247 Transportation Services 1 1 1 2 1 248 Communications 29 31 60 70 52 4449 Electric, Gas and Sanitary Services 7 8 6 7 5 750 Wholesale Trade - Durable Goods 8 6 14 12 11 951 Wholesale Trade - Nondurable Goods 7 6 9 8 4 452 Building Materials, Hardware, Garden Supply and
Mobile Home Dealers2 1
53 General Merchandise Stores 3 5 5 4 4 354 Food Stores 8 11 11 6 3 255 Automotive Dealers and Gasoline Service Stations 3 5 5 456 Apparel and Accessory Stores 3 4 8 6 5 557 Home Furniture, Furnishings and Equipment Stores 2 2 2 1 2 158 Eating and Drinking Places 3 5 5 2 4 359 Miscellaneous Retail 4 10 14 13 7 760 Depository Institutions 8 7 14 13 9 1061 Nondepository Credit Institutions 1 5 8 5 4 562 Security and Commodity Brokers, Dealers, Exchanges
and Services5 5 4 14 12 14
63 Insurance Carriers 10 11 14 9 12 1064 Insurance Agents, Brokers and Service 2 3 4 5 3 265 Real Estate 3 1 2 2 4 367 Holding and Other Investment Offices 2 3 5 4 2 370 Hotels, Rooming Houses, Camps and Other Lodging
Places1 2 3 4
72 Personal Services 2 4 3 2 173 Business Services 16 15 27 44 30 2875 Automotive Repair, Services and Parking 3 276 Miscellaneous Repair Services 178 Motion Pictures 9 8 8 10 6 679 Amusement and Recreation Services 3 3 6 10 8 680 Health Services 4 10 4 3 3 482 Educational Services 1 1 2 4 2 183 Social Services 187 Engineering, Accounting, Research, Management and
Related Services4 3 9 7 4 3
99 Nonclassifiable Establishments 1 1 4 3 1 1336 377 511 497 376 333
27
Table 1.4: Medians for Single and Dual Class Firms by YearThis table shows the medians, difference in medians, and the p-value for single and dual classfirms in the years 1998, 1992, 1996, 2000, 2004, and 2007. Dual-class firms are identified by theprocess described in Appendix C. Data comes from the CRSP/Compustat Fundamentals Annualdatabase.
1988 1992 1996 2000 2004 2007AssetsDual Class 139.5 204.3 265.5 537.0 739.5 984.1Single Class 55.3 71.7 108.6 188.6 311.9 483.7Difference 84.2 132.6 156.9 348.4 427.6 500.5p-value <.0001 <.0001 <.0001 <.0001 <.0001 <.0001
Debt / AssetsDual Class 19.5% 17.8% 17.2% 16.8% 15.9% 17.4%Single Class 12.7% 9.4% 6.7% 6.7% 7.5% 8.3%Difference 6.9% 8.3% 10.6% 10.1% 8.4% 9.1%p-value <.0001 <.0001 <.0001 <.0001 <.0001 <.0001
Sales GrowthDual Class 11.4% 5.8% 12.1% 11.8% 9.1% 7.7%Single Class 11.5% 7.3% 12.3% 12.8% 10.2% 9.0%Difference -0.1% -1.5% -0.2% -0.9% -1.1% -1.3%p-value 0.9895 0.1473 0.6845 0.9369 0.3086 0.0885
Tobin’s qDual Class 1.28 1.32 1.38 1.17 1.41 1.40Single Class 1.23 1.42 1.63 1.30 1.71 1.60Difference 0.05 -0.09 -0.24 -0.13 -0.30 -0.20p-value 0.7997 0.0008 <.0001 <.0001 <.0001 <.0001
Return on AssetsDual Class 3.08% 2.48% 2.81% 1.72% 3.16% 2.81%Single Class 2.41% 2.40% 2.03% 0.92% 1.81% 1.92%Difference 0.67% 0.08% 0.78% 0.80% 1.35% 0.89%p-value 0.0806 0.2189 0.0593 0.0002 0.0072 0.0897
Capital Expenditures / PPEDual Class 14.2% 10.5% 13.7% 13.0% 7.7% 10.5%Single Class 12.2% 10.6% 15.2% 14.7% 9.1% 11.5%Difference 2.0% -0.1% -1.5% -1.7% -1.3% -1.0%p-value 0.0338 0.2959 0.0649 0.241 0.0154 0.2502
R&D / SalesDual Class 2.61% 1.75% 1.58% 1.57% 1.64% 1.59%Single Class 3.51% 3.42% 4.68% 8.43% 7.26% 6.84%Difference -0.91% -1.67% -3.10% -6.86% -5.62% -5.26%p-value 0.2027 <.0001 <.0001 <.0001 <.0001 <.0001
Chapter 2
No More Share Classes:
A Study of U.S. Dual-Class Stock Unifications
2.1 Introduction
In a firm with a single class of common stock, conflicts of interests arise due to the separation
of ownership and control between the owner-manager of the firm and its outside stockholders
(Jensen and Meckling 1976). The owner-manager may shirk his responsibility to outside
stockholders by exerting less than maximum effort or by reaping pecuniary or non-pecuniary
benefits that are not in the best interests of outside stockholders. On the other hand, outside
stockholders may not monitor the firm, free riding on others. In order to mitigate the shirking
which arises from both parties due to the separation of ownership and control, Alchian and
Demsetz (1972) suggest owners maintain a bundle of five key rights and with those rights
transfer decision making/monitoring authority to a smaller group, the board of directors.
These five key rights are the right to residual cash flows, the right to alter the membership
of the board of directors (voting rights), the right to observe input behavior, the right to be
the central party common to all contracts with inputs, and the right to sell these rights. In
the typical single class stock structure, each share of stock holds the same right to residual
cash-flows and voting. However, under a dual class stock structure, these two key ownership
rights are unbundled. In some cases, one class may hold complete control of the voting rights,
whereas in others they may hold a disproportionate share of the vote. For example, Google’s
dual class structure consist of approximately 237 million class A shares, each with the right
to one vote, and approximately 77 million class B shares, each with the right to ten votes.1
1From Google’s 2008 Proxy statement (DEF-14A).
28
29
So while the class B shares hold only 25% of the cash-flow rights, they hold 76% of the voting
rights.
This unbundling of cash flow and voting rights further exacerbates the separation of
ownership and control problem. In a single class firm, there is a trade-off between increased
incentives which arise from ownership concentration (Jensen and Meckling (1976), Schleifer
and Vishny (1997)) and the risk of entrenchment (Stulz (1988), Morck, Shleifer, and Vishny
(1988)). In a dual class firm, the trade-offs are more obvious and extreme. Those with control
may become entrenched while at the same time have lower incentives due to reduced cash-flow
ownership (Claessens et al. (2002), Gompers, Ishii, and Metrick (2008)). This provides those
in control with increased opportunities to exploit (Gilson (1987)) and extract private benefits
from minority shareholders (Barclay and Holderness (1989), Nenova (2003)). The dual class
structure also acts as an effective antitakeover mechanism, preventing minority shareholders
from reaping sizable takeover premiums (Seligman (1986), Jarrell and Poulsen (1988), Gom-
pers, Ishii, and Metrick (2008)). In line with the entrenchment and private benefits extraction
theories, studies have found negative announcement effects at the implementation of the dual
class structure (Jarrell and Poulsen (1988)), negative operating performance after the imple-
mentation of the structure (Mikkelson and Partch (1984)), and negative firm value effects
due to the structure (Gompers, Ishii, and Metrick (2008), and citeasnounvillalonga:ffc).
On the other hand, the structure allows those with controlling interest and limited funds
to retain control while also accessing the equity markets for additional financing (DeAngelo
and DeAngelo (1985)). This may be especially beneficial for firms who require large amounts
of organization-specific human capital, whose projects are difficult for outsiders to value
due to high levels of information asymmetry, or for firms with high amenity potential, like
media outlets (Demsetz and Lehn (1985)). In line with this research, studies have found
positive announcement effects to the implementation of the structure (Partch (1987), Ang
and Megginson (1989), Cornett and Vetsuypens (1989)), positive industry-adjusted operating
performance after the structure’s implementation (Lehn, Netter, and Poulsen (1990)), and
30
positive long-term abnormal stock performance following the implementation (Dimitrov and
Jain (2006)).
Despite the mixed empirical results, many in the investment community voice their dis-
content with the dual class structure. Gary Hawton, chief executive of Meritus Mutual Funds,
states “It’s difficult for a board to say that they are being responsive to the needs of all share-
holders when they appear to bow down to the needs of the shareholder with the multiple
votes.”2 While discussing the pricing of Google’s IPO, Peter Chapman, senior vice-president
of TIAA-Cref, stated “there should be a substantial discount for corporate governance defi-
ciencies. This [dual class] structure effectively disenfranchises outside shareholders.”3
While the dual class structure potentially allows a family or institution to maintain control
of the firm indefinitely, there are occasions where dual class firms eliminate the disparate
voting rights and move to one vote per share. This recapitalization to a single share class is
commonly referred to as a“unification” in the literature.4 In this paper, I add to the literature
by using 61 American dual class share unifications to distinguish between the value recovery
and optimal structure hypotheses. Under the value recovery hypotheses, a firm eliminates
its dual class structure because the structure has reduced firm value (Villalong and Amit
(2008), and Gompers, Ishii, and Metrick (2008)) and has hampered operating performance
(Mikkelson and Partch (1984)). Under the optimal structure hypothesis, a firm eliminates its
dual class structure because the firm has become inherently different from other dual class
firms.
The rest of the paper is organized as follows. In Section 2, I review the extant liter-
ature on unifications and layout the predictions of the value recovery and optimal struc-
ture hypotheses. In Section 3, I discuss the unification sample, the control sample, and the
matching procedure. In section 4, I review the results and Section 5 concludes.
2“Dual class shares breed resentment”, Canadian Press, May 30, 2005.3“U.S. Fund Criticizes Google’s IPO Structure”, Financial Times, Simon London, May 4, 2004.4In this study, the term ‘recapitalization’ refers to the introduction of a dual class stock structure,
while a ‘unification’ refers to the elimination of a dual class stock structure.
31
2.2 Unifications: Extant Literature and Hypotheses
Like the 1920s, dual class firms are again receiving pressure to eliminate their unequal voting
policies.5 Institutional investors, unions, and large minority shareholders are calling on dual
class firms to move to one vote per share. The New York Times has been under pressure
from institutions, such as Morgan Stanley, to eliminate its dual class structure. Google, whose
recent IPO was one of the most anticipated in recent history, is criticized in the press for
having a dual class structure and was ranked lower than all S&P 500 firms for its corporate
governance structure by Institutional Shareholder Services.6 The dual class structure of Dow
Jones received criticism during the takeover bid by Rupert Murdoch because of the control
left in the hands of the Bancrofts, who held approximately 64% of the voting power and only
25% of the cash-flow rights.
Since 1992, at least 61 American dual class firms have chosen to eliminate their dual class
stock structures and move to one-vote per share.7 In a typical unification, each restricted and
superior voting share is exchanged for one share of the new common stock. In some cases,
the superior voting shares are exchanged at a higher rate than the restricted voting shares.8
For example, the superior voting shares (SVS) of Readers’ Digest Association received 1.22
shares of the new common stock, whereas the restricted voting shares (RVS) received only
1 share each. A special stockholder meeting is usually held after receiving board approval
and a proxy statement is issued outlining details of the planned unification. In addition,
the proxy usually outlines reasons why the firm is eliminating the structure. The two most
common reasons are to eliminate potential investor confusion (including calculation of market
capitalization), and to increase the investor base and liquidity of the firm’s shares. As an
5“Our Company Right or Wrong - Family Capitalism”, The Economist, March 17, 2007.“Class Struggle”, CFO Magazine, Andrew Osterland, October 1, 2001.
6Dow Jones Newswires - Google lands at bottom of ISS Governance Ranking, August 23, 2004.7In this study I do not include firms which moved to one share class automatically based on a
conversion of superior voting shares or a minimum threshold.8This occurs in approximately 20% of the sample
32
example, see Appendix A which outlines the time-line and key information in regards to a
unification at E-Z-EM.
The removal of the dual class structure leads to interesting questions. Since there are pri-
vate benefits from control, why would the majority vote holders be willing to give up control
in moving to one vote per share? If there are mixed results for the dual class implementa-
tion announcement, what is the stock market’s reaction to the unification announcement?
Considering dual class studies such as Dimitrov and Jain (2006), Lehn, Netter, and Poulsen
(1990), and Gompers, Ishii, and Metrick (2008), does the unification have an effect on firm
value and performance?
citeasnounang1989rvs, and Smart, Thirumali, and Zutter (2008) examine total announce-
ment effects (market capitalization) and find positive abnormal returns around the announce-
ment date.9 Dittman and Ulbricht (2008) examine the returns for the share classes, separately
and jointly, and find positive announcement returns of 5% (RVS) and 2.5% (SVS) for the
two-day event window (-1 to 0). In this study, I extend the literature by separately examining
the RVS and SVS announcement returns for American dual class unifications.
Hauser and Lauterbach (2001) examine the price of vote in Israeli unification, while
Bigelli, Mehrotra, and Rau (2008) look at 46 Italian unifications and suggests majority
shareholders take advantage of minority shareholders during the unification. Dittman and
Ulbricht (2008) also find ownership structure and changes in liquidity explain a signficant
portion of the cross-sectional variation in abnormal returns. Ehrhardt, Kuklinski, and Nowak
(2005) find an increase in share liquidity for German firms after the unification. In this study,
I extend the literature by examining how the unification affects share liquidity in American
firms.
Maury and Pajuste (2007) observe 105 European unifications to examine the determinants
and consequences of unifications. They find that firms with smaller wedges (voting - cash flow
rights), higher presence of financial investors, and higher frequency of cross-listing are more
9Ang and Megginson’s (1989) result may not be significant. The result is found in the text ofthe paper and the significance is never discussed.
33
apt to unify their shares. In this study, I perform the first American analysis of determinants
and consequences of unifications. For further detail on the prior literature, refer to Table 2.1.
The following two subsections outline the predictions of the value recovery and optimal
structure hypotheses.
2.2.1 Value Recovery Hypothesis
With a dual class structure in place, a family member or a member of management may
effectively maintain control of the firm while owning a small percentage of the ownership
rights. This concentration of control gives the holder the power to effectively veto any poten-
tial takeover or threat to management. Seligman (1986) states “Few takeover defenses are
more likely to be successful than dual class capitalization.” In fact, it was during the active
takeover market of the 1980s that the structure became more common. While the structure
guards against takeovers, it also prevents minority shareholders from reaping the benefits of
an average takeover premium of approximately 16-20%.10 Jarrell and Poulsen (1988) examine
94 dual class recapitalizations and find a significantly negative abnormal return around the
announcement of a dual class recapitalization. They attribute this result to the dual class
structures’ anti-takeover property.
In addition, this entrenchment allows the holder to influence decisions and exert control
in firms which would not be possible without the dual class structure. Barclay and Holderness
(1989) document these private benefits of control by examining premiums assigned in sales
of large blocks of stock. Masulis, Wang, and Xie (2008) identify four sources of extraction of
benefits in dual class firms that lead to depressed firm value. In a study of U.S. dual-class
firms, Gompers, Ishii, and Metrick (2008) find firm value decreases as the wedge between
control and cash-flow rights increases and citeasnounvillalonga:ffc find the dual class structure
has a negative impact on family firm values (industry-adjusted Tobin’s Q).
Based on the value recovery hypothesis, I expect:
10Andrade, Mitchell, and Stafford (2001) and Boone and Mulherin (2007)
34
• a positive stock price reaction to the unification announcement due to the removal
of the anti-takeover mechanism and the potential reduction in extraction of private
benefits of control (Masulis, Wang, and Xie (2008) and Gompers, Ishii, and Metrick
(2008))
• an increase in firm value and operating performance after the unification due to the
reduction in extraction of private benefits of control and the removal of the dual class
structure (Masulis, Wang, and Xie (2008), Gompers, Ishii, and Metrick (2008), and
citeasnounvillalonga:ffc).
2.2.2 Optimal Structure Hypothesis
The single class stock structure does not optimally fit the needs of every firm. In some firms,
the dual class structure allows management to make long-term investment decisions without
risk of scrutiny from outside monitors who have insufficient information. The structure allows
control to remain among the family or control group, while allowing for the inflow of addi-
tional capital to finance positive net present value projects. DeAngelo and DeAngelo (1985)
suggest the dual class structure is an “intermediate organizational form.” Certain firms which
require a certain level of control but who need additional external funding to finance positive
net present value projects may find the dual class structure optimal. In other words, there
are specific firm characteristics that optimally align the dual class structure with the needs
of the firm. In fact, when comparing firms who introduced a dual class structure to firms
who use leveraged buyouts, Lehn, Netter, and Poulsen (1990) find the firms that recapitalize
with dual classes have greater growth opportunities and routinely access the equity markets
after the recapitalization.
During a firm’s life cycle, firm characteristics can shift and the same firm who found the
dual class structure optimal at one point in time may later find the single share structure to
be optimal. In these cases, the company may decide it is in the best interest for shareholders
35
if the company unifies its existing two classes of stock. In effect, the costs of maintaining
control has now exceeded the benefits of retaining that control.
Under the optimal structure hypothesis, I expect:
• a positive initial stock price reaction to the unification announcement because the
unification is the result of the unexpected move toward optimality
• to find firm characteristics that are consistent among unifying firms and which set them
apart from non-unifying firms
2.3 Data and Sample Description
I identify 61 dual class class unifications by searching Lexis-Nexis, Factiva, and firm proxy
statements.11 I use search terms such as “dual class” and “single class of common stock” to
identify unifications. For the announcement date, I use the earliest article which discusses a
unification or potential unification. If an announcement article cannot be found, I use the
filing date of the first proxy statement that discusses the unification. I also collect share
class data and unification details by examining articles and proxy statements. Table 2.2 lists
the unifications by year and exchange. The unifications are evenly distributed between years
1992-2006 with a maximum of 9 in 1999. The majority of the sample firms are listed on
the NYSE (25), followed by Nasdaq (30), AMEX (5) and one over-the-counter. For each
unification, stock price data is collected for both the restricted and superior voting shares (if
traded publicly) from the Center for Research in Security Prices (CRSP). Accounting infor-
mation is collected from Compustat. Table 2.3 lists the number of unifications by two-digit
SIC code. The unifications are distributed across 35 two-digit SIC codes, with a maximum
of six unifications in the instrument and insurance carrier industries.
Table 2.4 lists statistical data on the unifications. For firms with both shares trading
publicly, the percentage market capitalization consisting of superior voting shares is 41.6%,
11A few proxy statements lists the names of other firms who have conducted unifications.
36
with a high of 81.7% and a low of 3.0%. The voting premium for superior voting shares the day
before the unification announcement averages 3.2%; however the range varies dramatically
from a maximum of 31.4% to a minimum of -9.4%. On average, the dual class structure is in
place 8.7 years before unification and it takes an average of 90 days for the unification to be
implemented after the announcement.12 Table 2.5 lists summary statistics for the sample of
firms in the study and their matched dual class counterparts. The medians are similar across
variables with the exception of shareholders. Unifying firms have a median of 2.3 million
shareholders while the matched dual class firms have a median of 0.6 million shareholders.
2.3.1 Universe of Dual Class Firms
For comparison, I identify a sample of dual class firms who did not unify their shares. The
most comprehensive dual class database to-date is the database created by Gompers, Ishii,
and Metrick (2008). Their sample comes from combining dual class data from Securities
Data Corporation (SDC), Compustat, the Center for Research in Security Prices (CRSP)
and the Investor Responsibility Research Center (IRRC). The sample runs from 1995-2002
and consists of 734 firms and 3,600 firm-years.
Because their database covers only years 1995-2002, it limits the ability to use their
database for matching the sample. As a solution, I have refined the CRSP-Compustat share
algorithm used by Gompers, Ishii, and Metrick (2008) and Zhang (2002) so that the time
frame may be extended while also limiting the pool of candidate firms. In their algorithm they
identify potential dual class firms if the difference in common shares outstanding between
the two databases is greater than 1%. This results in a candidate pool of more than 8,700
firms and 25,000 firm-years for the time frame 1976-2005. By refining the algorithm to include
filters which consider other stock measures and reporting errors in the CRSP and Compustat
data files, the revised algorithm reduces the pool of candidates by 45% to 2,770 firms and
13,580 firm-years at the same 1% level of difference.
12Of the 61 firms, 28 were dual class at IPO, 26 recapitalized to dual class, and 7 became dualclass through a dividend or spinoff.
37
For this study, I raise the threshold from 1% to 5% to further guarantee the firms which
the algorithm finds are in fact dual class. Using the algorithm and the increased threshold,
I identify 1,558 potential dual class firms and 8,780 firm-years for the period 1976-2006.
Table 5 shows the number of candidate firms by year along with a comparison of the firms
identified by Gompers, Ishii, and Metrick (2008). To test the revised algorithm, I compare
the candidate pool to the unifying firms in the study. The algorithm correctly identifies 50
out of 61 (82%) of the sample as dual class firms. As a further test of the algorithm, I examine
the 97 dual class firms identified in Table 9 of Jarrell and Poulsen (1988). I was able to locate
84 of these firms in Compustat and 68 (81%) are identified using the algorithm.
2.3.2 Matching Procedure
To compare the performance of unifying dual class firms to those who do not unify, I follow
Barber and Lyon (1996) and match on pre-event performance and industry. Using the control
sample, I first identify all firms with the same two-digit SIC code. Then I identify all firms
with plus or minus 10% return on assets (income before extraordinary items divided by book
value of assets) in the fiscal year prior to the unification are chosen. Lastly, I choose the firm
with performance closest to the unifying firm. Where no match is available, I filter firms first
on one-digit SIC code and then return on assets. If there is still no match available, I simply
match firms on return on assets, regardless of the firm’s industry. Once this procedure is
complete, I use firm proxy statements to verify the selected matching firms are indeed dual
class firms.
2.4 Results
2.4.1 Unification Announcement Event Study
Both the value recovery and optimal structure hypotheses predict a positive announcement to
the unification of the dual class structure; however, studies such as Dimitrov and Jain (2006)
find the dual class structure to be a value enhancing corporate initiative. If the structure is
38
value enhancing, then the removal of the structure may be a value decreasing event and lead
to a negative announcement effect. Using an event study analysis, I examine the unification
announcement effect to determine whether the results confirm the predictions of the value
recovery and optimal structure hypotheses.
From the sample of 61 unification announcements, I eliminate all unifications with con-
flicting events on the announcement date. For example, J.M. Smucker Company announced
on May 16, 2000 it would seek shareholder approval for a unification; however, the firm
simultaneously presented downward earnings guidance. After removing unifications with con-
flicting events, 36 announcements remain. To analyze the unification announcement effect, I
perform a standard event study analysis of the 36 announcements. I separately perform the
analysis on the restricted (RVS) and superior (SVS) voting shares. I use the market model
with the CRSP value-weighted market portfolio as the market proxy and days -250 to -6 as
the estimation period. As a robustness check, I also calculate net-of-market returns using
the CRSP value-weighted portfolio as a proxy for expected return. I examine the three-day
(-1 to +1) and five-day (-2 to +2) event windows to be consistent with previous studies.
Panel A of Table 2.7 contains the event study results for the restricted voting shares.
Of the 36 remaining announcements, the restricted voting share trades publicly in 32 firms.
At the announcement, the restricted voting shares have a three-day (-1 to +1) abnormal
return of 1.64%, significant at the 10% level. The three-day abnormal return is positive in
21 of the 32 firms and the net-of-market results confirm the results of the market model. For
the five-day event window (-2 to +2), the abnormal return increases to 1.85%, significant at
the 10% level. In the extant literature, two studies examine the unification announcement
effect separately for restricted and superior voting shares. Bigelli, Mehrota, and Rau (2008)
find three-day and five-day abnormal returns of 11.67% and 12.50% respectively for non-
voting shares in Italian unifications and Dittman and Ulbricht (2008) find a two-day (-1 to
0) abnormal return of 5.2% for non-voting shares in German unifications.
39
Panel B of Table 2.7 contains the results for the superior voting shares. Superior voting
shares trade publicly in 26 of the 36 announcements. The three-day (-1 to +1) abnormal
return for the superior voting shares is 1.15% and insignificant at the 10% level; however,
15 of the 26 are positive. For the five-day event window (-2 to +2), the abnormal return
increases to 1.73% and is not significant. The net-of-market return is 2.54% and significant
at the 10% level. Dittman and Ulbricht (2008) find a similar two-day (-1 to 0) abnormal
return of 2.5% for voting (superior) shares. In contrast, Bigelli, Mehrota, and Rau (2008)
find negative and significant abnormal returns of -1.56% (-1 to +1) and -1.94% (-2 to +2).
In order to examine the total announcement effect, I examine market capitalization. The
market capitalization results are shown in Panel C of Table 2.7. For the 36 firms, the total
three-day (-1 to +1) abnormal return is 1.30% and insignificant. The five-day (-2 to +2)
abnormal return is 2.23% and significant at the 5% level. In their German study, Dittman
and Ulbricht (2008) find a two-day (-1 to 0) abnormal return of 3.3% and Bigelli, Mehrota,
and Rau (2008) find an insignificant five-day (-2 to +2) return of 0.08% in their Italian
study.13 Similar to my result, Smart, Thirumalai, and Zutter (2008) find a significant five-day
abnormal return of 2.7% for market capitalization in a study of 37 American unifications.14
Both the value recovery and optimal structure hypotheses predict a positive announce-
ment effect to the elimination of the dual class structure. Based on the results of the event
study, I conclude there is a positive announcement effect to the unification and therefore find
support for both the value recovery and optimal structure hypotheses.
13In a British study, Ang and Megginson (1989) examine six unifications and find a mean two-day(-1 to 0) return of 0.65% (tests for significance are not reported).
14Smart, Thirumalai, and Zutter (2008) examine dual class initial public offerings from 1990 to1998. They find 37 unifications in their sample; however, no unifications are eliminated based onconflicting events.
40
2.4.2 Liquidity Analysis
A common reason given in firm proxy statements for unifying the dual class structure is the
expected increase in share liquidity (Maury and Pajuste (2007)).15 This reasoning implies the
dual class structure decreases share liquidity. Kim, Lin, Singh, and Yu (2007) confirm this
by finding increases in effective spreads and price impacts for both superior and restricted
voting shares after the dual class recapitalization. These increases in illiquidity can lead to
higher expected returns/cost of capital (Amihud and Mendelson (1986)). Based on Kim et
al. (2007) and firms commonly giving increased liquidity as a reason to unify their share
classes, I expect to find an increase in liquidity after the dual class unification.
To test this hypothesis, I use the bid-ask spread and Amihud (2002) measure. For each
measure, I examine the 50, 100, and 250-day window before the unification announcement
and the 50, 100, and 250-day window after the implementation of the single class structure.
The pre-announcement windows end on day -5, with day zero being the announcement day,
and the post-implementation windows begin on day 5, with day zero being the the first day
the single class is traded. Firms with share prices less than $5 are excluded.
Table 2.8 outlines the results of the liquidity analysis. Using the Amihud (2002) measure
of illiquidity, I find consistent reductions in the means and medians for both restricted and
superior voting shares across all time windows. For example, the pre-announcement mean
50-day Amihud measure for restricted voting shares is 1.367x10−6 and drops to 0.467x10−6
after implementation. The increase in liquidity is significant for both the restricted and
superior voting share windows. For the bid-ask spread, I find similar results. There is a
significant reduction in illiquidity across both classes for all time windows. For example, the
pre-announcement 100-day bid-ask spread for superior voting shares is 3.21% and drops to
1.69% after the implementation.
The results of the liquidity analysis show firms who unify their dual share classes sig-
nificantly increase the liquidity of their stock and confirms why companies commonly use
15See Appendix A - Sample Unification
41
liquidity as an explanation when moving to a single class of stock. In the same vein, Li,
Ortiz-Molina, and Zhao (2008) find institutional investment increases after the unification
and Dittman and Ulbricht (2008) find liquidity helps explain the variation in abnormal
returns during the unification announcement. Also, Ehrhardt, Kuklinski, and Nowak (2005)
find a significant reduction in bid-ask spreads following German unifications.
2.4.3 Unification Determinants
To distinguish between the value recovery and optimal structure hypotheses, I conduct a
probit analysis to examine the determinants of the choice to unify share classes. The primary
data sample used in the probit analysis comes from Gompers, Ishii, and Metrick (2008)
(GIM). Their dataset provides complete share class information (voting rights, dividend
rights, volume, and permnos) for dual class firms from 1995 to 2002.16 Of the 61 unifications
included in the study, 33 are located in the GIM dataset. Firms who unify their share classes
are assigned a one for the dependent variable in the year prior to their unification and then
drop out of the sample.
The following variables are included in the analysis:
Control wedge is the difference between the voting (control) percentage and cash-flow
(ownership) percentage owned by the officers and directors of the firm. The control wedge
measures the size of the separation between voting and ownership rights in the firm. Gompers,
Ishii, and Metrick (2008) find firm value (Tobin’s Q) decreases as the control wedge increases.
Masulis, Wang, and Xie (2008) find evidence managers of dual class firms with greater control
wedges are more likely to extract private benefits of control. Maury and Pajuste (2007) find
dual class firms with smaller control wedges are more likely to unify their share classes.
Control percentage and ownership percentage are the individual components of control wedge.
Based on prior literature, such as Gompers, Ishii, and Metrick (2008) and Maury and Pajuste
16This information has not been collected for the universe of potential dual class firms discussedin section 3.1.
42
(2007), I expect dual class firms with lower control percentages to be more likely to unify
their share classes.
Capital expenditures are the amount of capital expenditures divided by the previous year’s
net property, plant, and equipment. I use capital expenditures as a proxy for growth in the
firm. Lehn, Netter, and Poulsen (1990) find firms who move to the dual class structure have
higher capital expenditures than those who have a leveraged buyout.
Net income is the firm’s income before extraordinary items divided by sales. Claessens
et al. (2002) suggest the expropriation of minority shareholders and extraction of private
benefits of control increase with the dual class structure. Masulis, Wang, and Xie (2008)
identify four methods in which dual class managers are able to extract private benefits of
control from restricted vote holders. Thus, I include net income to examine the effects of
private benefit extraction prior to the unification.
Equity issue proceeds measures the net equity proceeds divided by total shareholder’s
equity. Maury and Pajuste (2007) find dual class firms who unify have higher equity issuance
in the fiscal year before unification than do non-unifying dual class firms. On the other hand,
Amoako-Adu and Smith (2001) identify “increase investor appeal prior to seasoned offering”
as one of the most common reasons given why dual class firms on the Toronto Stock Exchange
choose to unify their classes of stock, which may imply firms delay equity issuance until after
the unification.
Amihud measure is the 250-day weighted average of the Amihud (2002) measure of illiq-
uidity. Firms commonly claim increasing liquidity as a reason for eliminating the dual class
structure. The illiquidity of the firm’s shares can increase the cost of capital (Amihud and
Mendelson (1986)). Kim et al. (2007) find share liquidity decreases at the introduction of the
dual class structure and in the previous section I find a significant increase in share liquidity
after the implementation of the single class structure.
43
Leverage and firm size are also included and firm size is measured as the natural logarithm
of the book value of assets. Year dummies are included in the analysis and standard errors
are corrected for heteroskedasticity.
Table 2.9 outlines the results of the probit analysis. Because of data limitations, the
final sample includes 1,639 firm years and 27 unifications. Model 1 and 2 are similar except
the control wedge is separated between control (voting) and ownership in Model 2. The
control wedge in Model 1 is negative and significant at the 5% level and corresponds to
previous literature (Maury and Pajuste (2007), Dittman and Ulbricht (2008), and Ehrhardt,
Kuklinski, and Nowak (2005)). In Model 2, the wedge is separated and the control percentage
is negative and significant at the 5% level while the ownership percentage is positive but not
significant. These results show dual class firms with smaller wedges and less voting control
are more likely to unify their share classes.
In Models 1 and 2, the Amihud (2002) measure is positive and significant at the 1% level.
In the previous section, I show an increase in share liquidity after the unification and Kim et
al. (2007) shows the dual class structure reduces share liquidity. The probit analysis shows
dual class firms with higher illiquidity are more likely to unify their share classes.
For net income, I find an insignificant relationship in both models, so firms who unify
are no more likely to have better or worse earnings than those who remain dual class. This
implies firms who unify have no more or less private benefit extraction than non-unifying
dual class firms. For both models, equity issues are insignificant. This contrasts with findings
by Maury and Pajuste (2007) and Ehrhardt, Kuklinski, and Nowak (2005) who find dual
class firms with more equity issues are more likely to unify.
In Model 1 and 2, capital expenditures are positive and significant at the 5% level. These
results are evidence dual class firms who are investing more in property, plant, and equipment
are more likely to unify their share classes. In addition, the leverage variable is positive and
significant for both models, implying dual class firms with higher leverage are more likely to
unify their share classes.
44
In summary, the probit analysis finds firms with smaller control wedges, and higher
illiquidity, capital expenditures, and leverage are more likely to unify their share classes. I
find these results to be in line with the optimal structure hypothesis, which predicts there
are key characteristics in unifying firms, such as high illiquidity, capital expenditures, and
leverage, that distinguishes them from other dual class firms and leads them to unify their
share classes.
2.4.4 Firm Value and Operating Performance Post-Unification
In order to further test the value recovery and optimal structure hypothesis, I use the fol-
lowing specification (Pagano, Panetta, and Zingales 1998) to examine firm value, perfor-
mance, and other firm characteristics after the unification:
yit = α +3∑
j=0
βjUNIt−j + β4UNIt−n + ui + dt + εit
where ui and dt are firm-specific and fiscal-year specific effects. UNIt−j are dummy variables
equal to one if year t− j was the unification year, UNIt−n is a dummy variable equal to one
if the unification took place more than three years before. By using a fixed-effects model,
I use each company before the unification as a control for itself after the unification. Table
2.10 outlines the results.
The first six rows of Table 2.10 show the post-unification yearly effects on Tobin’s Q. For
all three measurements, there is not a significant change in any one year and the accumulated
F -test for years zero to two is insignificant as well. These results contrast those found in
Maury and Pajuste (2007), who find a positive and significant increase in industry-adjusted
market-to-book in year 0 and +1. In addition, the results contrast studies that show the
dual class structure has a negative effect on firm value, such as Gompers, Ishii, and Metrick
(2008) and citeasnounvillalonga:ffc.
For operating performance, the return on assets results are all insignificant. Unadjusted
net income to sales is insignificant for all years. The dual matched net income results show a
significant increase in year two and the three year F -test rejects the null of no change. The
45
industry-adjusted figures are significant in year zero and two and the F -test reject the null
of no change for the year zero to two time frame.
I also examine other variables after the unification. Net stock issuance is negative in years
zero and one and insignificant for the three year test period. On the other hand, leverage is
positive and significant in years one and two, as well as the F -test. Taken together, these
results show that unifying firms are not issuing new equity after unification. Sales growth
is negative in all years except year three and the F -test is significant at the 5% level. In
contrast, Maury and Pajuste (2007) find a positive and significant increase in sales growth
after the unification.
In summary, the results show no significant change in firm value and conflicting results
about operating performance after the unification. These results show no evidence of a
recovery in value or performance by the elimination of any private benefits of control and
are evidence against the value recovery hypothesis.
2.4.5 Firm Events after Unification
Amoako-Adu and Smith (2001) cite “facilitate sale of control block” as one of the most
common reasons Canadian firms eliminate their dual class structure. So, as an alternative
hypothesis, firms may leave the dual class structure simply to allow the control holder to
sell out. In order to test this hypothesis among American dual class firms, I use Compustat,
Hoover’s, and Factiva to examine the current status of unifying firms and their matched
counterparts. If a firm has merged or gone bankrupt, I obtain the date of the transac-
tion. Using this information, I examine post-unification firm events (control changes). Table
2.11 reports the results of the analysis. For all time periods, there is no significant differ-
ence between the frequency of mergers/acquisitions in unifying and non-unifying firms. For
example after three years, 85% of the matched firms remain publicly traded and 83% of the
unifying firms remain so. In a similar manner, Maury and Pajuste (2007) analyze ownership
changes post-unification and find the unification of the dual class structure does not lead
46
to an exit by the controlling shareholder. I find this evidence does not support the “sale
of control block” hypothesis since non-unifying firms are just as likely to undergo control
changes as are unifying dual class firms.
2.5 Conclusion
The dual class structure allows for a separation between two of the key rights Alchian and
Demsetz (1972) identified as necessary for owners of a modern corporation. This separation
of voting and cash-flow rights has again become an important issue in the investment com-
munity as evidenced by firms such as the New York Times, Google, and Dow Jones. In this
study, I use 61 American dual class unifications to distinguish between the value recovery and
optimal structure hypotheses. In line with both hypotheses, I find a positive and significant
market reaction to the elimination of the dual class structure. I also add to the literature
by demonstrating a significant increase in liquidity for American firms who leave the dual
class structure. To separate the hypotheses, I perform a probit analysis and find unifying
firms have lower control wedges, higher leverage and capital expenditures, and higher levels
of illiquidity. As further evidence against the value recovery hypothesis, I find no significant
change in firm value and conflicting operating performance results after the elimination of
the structure. In addition, I find unifying firms are no more likely to be acquired or taken
private than their dual class counterparts.
47
Tab
le2.
1:D
ual
Cla
ssU
nifi
cati
onStu
die
s
Sam
ple
Sam
ple
Sam
ple
Oth
erS
tud
yD
escr
ipti
on
Per
iod
Siz
eIn
foA
ng
an
dM
eggin
son
(1989)
Bri
tish
Fir
ms
1955-1
982
6F
ind
exce
ssre
turn
sin
an
nou
nce
men
tm
onth
are
insi
gn
if-
icant,
but
do
find
asi
gnifi
cant
-23.4
5%
cum
ula
tive
exce
ssre
turn
du
rin
gth
e12
month
saft
eran
nou
nce
men
t.A
moako-A
du
an
dSm
ith
(2001)
Canad
ian
Fir
ms
1979-1
998
56
Exam
ine
the
reaso
ns
why
firm
su
nif
yst
ock
class
es.
The
thre
em
ost
com
mon
reaso
ns
are
(1)
un
ifica
tion
requ
ired
as
part
of
deb
tre
stru
cturi
ng,
(2)
faci
lita
tesa
leof
contr
ol
blo
ck,
an
d(3
)in
crea
sein
ves
tor
app
eal
pri
or
tose
aso
ned
off
erin
g.
Hau
ser
an
dL
au
terb
ach
(2004)
Isra
eli
Fir
ms
1990-2
000
84
Exam
ine
the
pri
ceof
vote
.F
ind
the
pri
ceof
vote
incr
ease
sw
ith
the
per
centa
ge
vote
lost
by
the
ma
jori
tysh
are
hold
ers,
ish
igh
erin
fam
ily-c
ontr
olled
firm
s,d
ecre
ase
sw
ith
inst
itu
-ti
on
al
inves
tor
hold
ings,
and
issi
milar
toth
ep
rice
of
vote
imp
lici
tin
the
mark
etp
rice
sof
stock
s.E
hrh
ard
t,K
uklin
ski,
an
dN
ow
ak
(2005)
Ger
man
Fir
ms
1997-2
003
43
Fin
dth
eu
nifi
cati
on
ofdu
alcl
ass
share
sto
be
stri
ctly
share
-h
old
ervalu
ein
crea
sing.
Fin
dd
ual
class
firm
sre
du
ceco
stof
cap
ital
thro
ugh
the
un
ifica
tion
du
eto
incr
ease
sin
firm
valu
eas
wel
las
red
uct
ions
inb
id-a
sksp
reads.
Mau
ryan
dP
aju
ste
(2007)
Conti
nen
tal
Eu
rop
e1996-2
002
108
Exam
ines
the
det
erm
inants
an
dco
nse
quen
ces
of
un
ifica
-ti
on
s.F
ind
sfi
rms
that
un
ify
have
low
erse
para
tion
bet
wee
nvoti
ng
an
dca
shfl
ow
rights
,h
igher
pre
sence
of
fin
an
-ci
al
inves
tors
,and
hig
her
freq
uen
cyof
cross
-lis
tin
gin
the
U.S
.F
ind
no
diff
eren
cein
ex-p
ost
sale
sgro
wth
and
cap-
ital
exp
end
itu
res
bet
wee
nun
ifyin
gfi
rms
and
those
that
rem
ain
edd
ual
class
.B
igel
li,
Meh
rotr
a,
an
dR
au
(2008)
Italian
Fir
ms
1974-2
005
46
Fin
dm
ajo
rity
share
hold
ers
hed
ge
or
take
advanta
ge
of
un
i-fi
cati
on
sby
engagin
gin
act
ivit
ies
month
sb
efore
the
un
i-fi
cati
on
dec
isio
n.
Fin
dth
ep
rice
of
voti
ng
share
sd
ropp
edfr
om
-4.2
6%
to-1
0.4
1%
at
un
ifica
tion
an
nou
nce
men
t.D
ittm
an
&U
lbri
cht
(2008)
Ger
man
Fir
ms
1990-2
001
29
Fin
dow
ner
ship
stru
ctu
rean
dch
an
ges
inliqu
idit
yex
pla
ina
sign
ifica
nt
part
of
the
cross
-sec
tion
al
vari
ati
on
inab
norm
al
retu
rns.
Als
o,
they
find
firm
sare
more
likel
yto
un
ify
wh
enth
eco
ntr
ollin
gsh
are
hold
erlo
ses
litt
levoti
ng
pow
erand
the
firm
isfin
anci
ally
con
stra
ined
.O
ften
,fi
rms
issu
ead
dit
ional
share
saft
erth
eu
nifi
cati
on
.L
i,O
rtiz
-Molina
and
Zhao
(2008)
Am
eric
an
Fir
ms
1995-2
002
79
Exam
ine
inst
ituti
onal
ow
ner
ship
ind
ual
class
firm
s.F
ind
inst
itu
tion
al
ow
ner
ship
issi
gnifi
cantl
ylo
wer
ind
ual
class
firm
sth
an
insi
ngle
-cla
ssfi
rms.
Follow
ing
un
ifica
tion
,th
eyfi
nd
inst
itu
tion
al
inves
tors
incr
ease
thei
rsh
are
hold
ings.
Sm
art
,T
hir
um
ali,
an
dZ
utt
er(2
008)
Am
eric
an
Fir
ms
1990-1
998
37
Fin
dsi
gn
ifica
ntl
yp
osi
tive
elev
enday
(-5to
+5)
ab
norm
al
retu
rnof
5.2
%.
48
Table 2.2: Distribution of unification announcements by year and exchangeThis table reports the distribution of unification announcements by year and exchange of a sampleof 61 dual class firms that moved to a single class of common stock during the 1992-2006 timeperiod. Unifications were identified by examining news articles from Lexis-Nexis and Factiva.
Observations by year of announcement Observations by exchange listingYear Announcements NYSE NASDAQ AMEX OTC1992 4 1 31993 3 2 119941995 2 1 11996 4 2 21997 2 1 11998 5 1 41999 9 1 6 22000 6 3 32001 5 4 12002 6 3 2 12003 7 4 32004 2 22005 4 2 22006 2 2Total 61 25 30 5 1
49
Table 2.3: Number of unifications by two-digit SIC codeThis table illustrates the two digit SIC code industry distribution for the 61 unifications identifiedin this study. Unifications were identified by examining news articles from Lexis-Nexis and Factiva.
SIC Code Industry Description No. of Unifications10 Metal Mining 113 Oil and Gas Extraction 415 Building Construction 220 Food and Kindred Products 126 Paper and Allied Products 127 Printing and Publishing 128 Chemicals and Allied Products 129 Petroleum Refining And Related Industries 132 Stone, Clay, Glass, And Concrete Products 133 Primary Metal Industries 134 Fabricated Metal Products 135 Industrial Machinery And Equipment 236 Electronic And Other Electrical Equipment 437 Transportation Equipment 138 Instruments and Related Products 640 Railroad Transportation 141 Local And Suburban Transportation 142 Motor Freight Transportation and Warehousing 145 Transportation By Air 148 Communications 150 Wholesale Trade-durable Goods 251 Wholesale Trade-non-durable Goods 154 Food Stores 158 Eating And Drinking Places 259 Miscellaneous Retail 260 Depository Institutions 262 Security And Commodity Brokers 163 Insurance Carriers 665 Real Estate 167 Holding And Other Investment Offices 173 Business Services 378 Motion Pictures 180 Health Services 182 Educational Services 187 Engineering and Management Services 3
50
Table 2.4: Unification StatisticsThis table presents relevant data on unifications. Panel A presents descriptive statistics on variousunification information. SVS % of market capitalization is the percentage of market capitalizationfrom superior voting shares on the day prior to the unification announcement for firms with bothclasses trading publicly. SVS Price Premium or voting premium is the superior voting share pricepremium on the day before the announcement and is calculated as follows:
PSV S,−1 − PRV S,−1
PRV S,−1
Length dual structure in place is the number of years the dual class stock structure was in placebefore the unification announcement. Time to implementation is the number of days between theunification announcement until the single class of stock began to trade. Panel B lists the tradingarrangement for the unification firms. Data comes from CRSP, Compustat, and firm proxies.
Panel A: Unification Data
StandardObs Mean Median Deviation Max. Min.
SVS % of Market Capitalization 39 41.6% 45.5% 19.5% 81.7% 3.0%SVS Price Premium (Voting Premium) 39 3.2% 0.0% 9.4% 31.4% -9.4%Length Dual Structure in Place (Years) 61 8.7 7.1 6.4 32.9 0.1Time to Implementation (Days) 61 90 77 70 393 0
Panel B: Pre-Unification Trading
ObsBoth Share Classes Trade Publicly 39Only Superior Voting Shares Trade Publicly 5Only Restricted Voting Shares Trade Publicly 17Total 61
51
Table 2.5: Summary StatisticsThis table presents descriptive statistics for 61 unifying firms and their matched counterparts forthe fiscal year prior to the unification announcement. Accounting data comes from Compustat.Total assets is the total assets of the firm in millions (data6). Sales is the company’s sales (data12).Leverage is long-term debt (data9) divided by the firm’s total assets (data6). Research anddevelopment is the firm’s research and development expenses (data46) divided by the firm’s sales(data12). Return on assets is the firm’s income before extraordinary items (data18) divided bythe firm’s total book value of assets (data6). Net income is the firm’s income before extraordinaryitems (data18) divided by the firm’s sales (data12). Return on equity is the firm’s income beforeextraordinary items (data18) divided by the firm’s book value of equity (data11). Tobin’s Q isdefined as the ratio of the firm’s market value to the firm’s asset replacement cost. Market value isproxied by the market value of common shares outstanding (data25*data199) plus preferred stock(data10) plus debt (data9+data5). Replacement cost is proxied by the book value of assets (data6).Sales growth is the current year’s sales less the previous year’s sales divided by the previous year’ssales. Capital expenditures is the firm’s capital expenditures from the statement of cash flows(data128) divided by the previous year’s net property, plant, and equipment (data8). Cash is thefirm’s cash and short-term investments (data1) divided by the firm’s assets (data6). CEO Pay isthe CEO’s total pay in millions (TDC1 taken from Execucomp). Employees is the number of firmemployees in thousands (data29). Stockholders is the number of stockholders in millions (data100).
Control firms come from a dual class firm dataset identified by using CRSP and Compustat. Firmsare first matched on two-digit SIC code and +-10% return on assets. If no match is available, firmsare matched based on one-digit SIC code and +-10% return on assets. If there is still no matchavailable, firms are matched on return on assets.
Sample Firms Control FirmsWilcoxon
N Mean Median N Mean Median Test StatisticTotal assets ($000,000) 61 2,027.6 431.7 61 1,928.3 412.4 (0.59)Sales ($000,000) 61 2,028.7 321.6 61 1,266.2 216.2 (1.23)Leverage 61 20.1% 18.5% 61 18.1% 13.7% (0.85)Research and development 23 8.8% 3.5% 24 6.8% 1.7% 0.89Return on assets 61 -3.0% 2.4% 61 -2.0% 2.2% (0.01)Net Income 61 -5.5% 4.1% 60 -11.8% 4.7% 0.44Return on equity 61 -6.3% 8.7% 61 -368.4% 8.0% (0.12)Tobin’s Q 51 1.89 1.36 51 1.77 1.21 (0.28)Sales growth 61 16.0% 7.8% 61 18.1% 9.1% 0.10Capital expenditures 57 34.1% 17.8% 58 44.6% 17.2% (0.50)Cash 61 15.4% 7.5% 61 14.1% 6.2% (1.09)CEO Pay (Millions) 25 4,516.8 2,061.2 15 7,471.7 2,146.7 (0.11)Employees (Thousands) 59 10.4 1.7 58 5.1 1.2 (0.97)Stockholders (Millions) 59 91.0 2.3 57 7.5 0.6 (3.41)***
52
Table 2.6: Universe of Potential Dual Class FirmsThis table demonstrates the number of potential dual class firms identified by the algorithm. Italso presents the number of firms identified as dual class by Gompers, Ishii, and Metrick (2008).
Year Algorithm GIM (2008)1976 551977 801978 781979 781980 931981 951982 1061983 1401984 1681985 2001986 2661987 3421988 3731989 3491990 3771991 3931992 3581993 3431994 3661995 382 4001996 420 4441997 483 4851998 535 5041999 512 4892000 439 4822001 370 4342002 332 3622003 3072004 2722005 2582006 210
53
Table 2.7: Unification Event Study ResultsThis table outlines the event study results of the unification of dual class shares. The market modelis estimated using CRSP’s value-weighted market portfolio as the market proxy and days -250 to-6 as the estimation period. The net of market method also uses CRSP’s value-weighted marketportfolio as the market proxy. The change in market capitalization was calculated as follows:
(PRV S,+t ∗ SHRSRV S,+t) + (PSV S,+t ∗ SHRSSV S,+t)(PRV S,−t ∗ SHRSRV S,−t) + (PSV S,−t ∗ SHRSSV S,−t)
− 1
where P is price, SHRS is shares outstanding, RV S represents the restricted voting class andSV S the superior voting class, +t is the end of the event window and −t is the beginning of theevent window. If only one share class trades publicly, the change in market capitalization is basedsolely on the one publicly traded class. Panel A lists the event study results for the restrictedvoting share class. Panel B lists the event study results for the superior voting share class. PanelC shows the change in total market capitalization (restricted and superior voting shares) for thespecified event window. *, **, and *** denote the results are significantly different from zero atthe 10%, 5%, and 1% levels, respectively.
Panel A: Restricted Voting Shares
Market Model Net of MarketMean Median % Mean Median %
Window N CAR p-value CAR Positive CAR p-value CAR Positive-1 to +1 32 1.64%* 0.075 0.60% 66% 2.21%** 0.031 0.48% 56%-2 to +2 32 1.85%* 0.097 1.95% 63% 2.56%** 0.022 2.33% 66%
Panel B: Superior Voting Shares
Market Model Net of MarketMean Median % Mean Median %
Window N CAR p-value CAR Positive CAR p-value CAR Positive-1 to +1 26 1.15% 0.375 0.70% 58% 1.81% 0.143 1.65% 62%-2 to +2 26 1.73% 0.263 1.74% 58% 2.54%* 0.079 2.69% 65%
Panel C: Change in Total Market Capitalization
Mean Median %Window N Change p-value Change Positive-1 to +1 36 1.30% 0.146 0.60% 58%-2 to +2 36 2.23%** 0.013 1.85% 69%
54
Table 2.8: Liquidity Around UnificationThis table shows the liquidity of shares for unifying firms before the unification announcement andafter the implementation of the unification. Data is collected from CRSP. Firms with share pricesless than $5 are excluded. The examination windows prior to the unification announcement endson day -5 and the window for after the implementation begins on day +5. For the Wilcoxon teststatistic both the RVS and SVS liquidity measures are compared to the same post implementationmeasure. Panel A outlines the Amihud (2002) measure of illiquidity.
Amihud’s Measure =|daily return|
dollar trading volume
Amihud measures are 10−6. Panel B outlines the bid-ask spread. *, **, and *** denote the resultsare significantly different from zero at the 10%, 5%, and 1% levels, respectively.
Prior to Announcement Post ImplementationWilcoxon
Window Class N Mean Median Std Dev N Mean Median Std Dev Test StatisticPanel A: Amihud Measure50 days RVS 41 1.367 0.059 4.297 46 0.467 0.010 1.454 2.01**
SVS 37 1.383 0.070 2.704 2.56**
100 days RVS 41 1.225 0.061 3.137 46 0.584 0.011 1.791 2.10**SVS 37 1.459 0.109 3.182 2.66***
250 days RVS 44 1.049 0.089 2.188 49 0.648 0.017 2.089 2.26**SVS 37 1.456 0.106 3.091 2.48**
Panel B: Bid-Ask Spread50 days RVS 41 3.00% 1.94% 3.86% 46 1.69% 0.90% 2.37% 2.13**
SVS 36 3.44% 2.06% 3.46% 3.16***
100 days RVS 41 3.20% 1.76% 4.22% 46 1.69% 0.92% 2.44% 2.32**SVS 36 3.21% 2.00% 3.23% 3.28***
250 days RVS 44 3.08% 2.12% 3.44% 49 1.86% 1.13% 2.58% 2.55**SVS 36 3.27% 2.40% 3.22% 3.25***
55
Table 2.9: Unification DeterminantsProbit models of the determinants of the decision to unify share classes. The primary dataset isbased on Gompers, Ishii, and Metrick (GIM) (2008). Accounting data is from Compustat and runsfrom 1995 - 2002. Firms who unify their share classes are assigned a one for the dependent variablein the year prior to their unification and then drop out of the sample. Independent variablesare as follows: Control wedge is the difference between the voting percentage and cash-flowpercentage owned by the officers and directors of the firm. Control percentage is the percent ofthe vote the officers and directors of the firm control. Ownership percentage is the percent of thetotal shares of the firm owned by the officers and directors of the firm. Amihud Measure is thefirm’s 250-day weighted-average (between RVS and SVS) Amihud measure. Net income is thefirm’s earnings (data18) divided by sales (data12). Equity issues is the firm equity issue proceeds(data108-data115) divided by the book value of equity (data11). Capital expenditures are the firm’scapital expenditures (data128) divided by lagged net property, plant, and equipment (data8).Leverage is the sum of long-term (data9) and current debt (data5) divided by the book value ofassets (data6). Firm size is the natural logarithm of the book value of assets (data6). Financialvariables are winsorized at the 1st and 99th percentiles. z-test statistics are reported in parenthesisbelow the coefficient estimates. Standard errors are corrected for heteroskedasticity. *, **, and ***denote the results are significantly different from zero at the 10%, 5%, and 1% levels, respectively.
Model 1 2Control Wedge -1.422**
(2.50)Control Percentage -1.370***
(2.68)Ownership Percentage 0.782
(1.06)Amihud Measure 0.058*** 0.065***
(2.63) (2.89)Net Income 0.519 0.501
(1.12) (1.21)Equity Issues -0.032 -0.032
(0.20) (0.22)Capital Expenditures 0.344** 0.354**
(2.46) (2.50)Leverage 0.640** 0.691**
(2.27) (2.41)Firm Size 0.050 0.035
(0.83) (0.59)
Log-Likelihood -125.05 -124.12McFadden Pseudo R2 9.2% 9.8%No. of Observations 1,639 1,639
56
Table 2.10: Post-Unification AnalysisFor each of the variables listed I estimate the following specification:
yit = α+3∑
j=0
βjUNIt−j + β4UNIt−n + ui + dt + εit
where ui and dt are firm-specific and fiscal-year specific effects. UNIt−j are dummy variables equalto one if year t − j was the unification year, UNIt−n is a dummy variable equal to one if theunification took place more than three years before. By using a fixed effects model I am using eachcompany before the unification as a control for itself after the unification. The table only reports thecoefficients on the UNI dummy variables. Tobin’s Q is defined as the ratio of the firm’s market valueto the firm’s asset replacement cost. Market value is proxied by the market value of common sharesoutstanding (data25*data199) plus preferred stock (data10) plus debt (data9+data5). Replacementcost is proxied by the book value of assets (data6). Return on assets is the firm’s income (data18)divided by the book value of assets (data6). Net income is the firm’s income before extraordinaryitems (data18) divided by the firm’s sales (data12). Capital expenditures is the capital expendituresdivided by last year’s net property, plant, and equipment. Sales growth is the percentage changein sales (data6) from the previous year. Leverage is long-term debt (data9) divided by the firm’stotal assets (data6). Net stock issuance is the net of stock sales (data108) and stock purchases(data115) divided by the previous year’s stockholder equity (data11). Payout is dividends (data21)divided by earnings (data18). Financial variables are winsorized at the 1st and 99th percentiles.Heteroskedasticity robust standard errors are reported in parenthesis. *, **, and *** denote theresults are significantly different from zero at the 10%, 5%, and 1% levels, respectively. The secondto the last column reports the p-value of an F -test of the hypothesis that the sum of the coefficientsfor dummies for year zero to two are equal to zero. The last column reports the p-value of an F -testof the hypothesis that the sum of the coefficients of all post-unification dummies are equal to zero.
Firm Year Year Year Year Year F-testVariable Years 0 +1 +2 +3 >3 Years 0-2Tobin’s Q 596 0.134 0.021 0.085 0.089 0.134 0.550
(0.161) (0.163) (0.169) (0.206) (0.204)Dual Matched Tobin’s Q 550 0.195 -0.014 -0.038 -0.054 -0.019 0.757
(0.191) (0.181) (0.198) (0.245) (0.242)Industry Adjusted Tobin’s Q 596 0.115 -0.059 -0.020 -0.021 -0.003 0.929
(0.177) (0.168) (0.172) (0.208) (0.204)Return on Assets 730 -0.005 -0.005 -0.013 -0.024 -0.043 0.590
(0.017) (0.019) (0.017) (0.022) (0.028)Dual Matched Return on Assets 690 0.008 0.014 0.005 -0.002 -0.026 0.592
(0.020) (0.022) (0.020) (0.024) (0.030)Ind Adjusted Return on Assets 730 0.005 0.008 0.011 -0.011 -0.035 0.582
(0.017) (0.018) (0.017) (0.021) (0.027)Net Income 730 0.033 0.029 0.023 0.023 0.023 0.246
(0.030) (0.031) (0.029) (0.033) (0.046)Dual Matched Net Income 690 0.121 0.123 0.264* 0.151 0.132 0.057*
(0.099) (0.084) (0.158) (0.100) (0.139)Industry Adjusted Net Income 730 0.057* 0.048 0.059* 0.028 0.008 0.034**
(0.033) (0.032) (0.032) (0.034) (0.044)Capital Expenditures 674 -0.025 -0.062 -0.069 -0.043 -0.064 0.146
(0.038) (0.052) (0.050) (0.060) (0.504)Sales Growth 710 -0.052 -0.094* -0.125* 0.000 -0.065 0.034**
(0.039) (0.052) (0.073) (0.076) (0.065)Leverage 728 0.028 0.042* 0.080** 0.038 0.066 0.004***
(0.019) (0.022) (0.029) (0.028) (0.030)Net Stock Issuance 628 -0.008 -0.111 0.182 0.217 0.099 0.921
(0.242) (0.196) (0.340) (0.365) (0.226)Payout 726 -0.007 -0.015 0.075 -0.038 -0.048 0.657
(0.052) (0.038) (0.063) (0.058) (0.062)
57
Table 2.11: Firm events post-unificationThis table illustrates the frequency in which major firm events occur for the 61 unifying firms andtheir matched dual class counterparts. Panel A outlines the major firm events for those firms whounified their share classes. For example within 5 years of the unification, 33% of firms had beenmerged, acquired, or taken private. Panel B outlines the major firm events which occurred for thematching dual class firms. Year 0 corresponds to the year the sample firm unified. Data comesfrom Compustat, Hoover’s, and Factiva
Panel A: Unifying Firms
+1 year +3 years +5 years +7 years +10 yearsN % N % N % N % N %
Remain publicly traded 59 97% 45 83% 31 67% 18 53% 9 60%Merger/Acquisition/Private 2 3% 9 17% 15 33% 16 47% 6 40%
BankruptTotal 61 54 46 34 15
Panel B: Matched Dual-Class Firms
+1 year +3 years +5 years +7 years +10 yearsN % N % N % N % N %
Remain publicly traded 55 90% 46 85% 33 72% 17 50% 8 53%Merger/Acquisition/Private 6 10% 8 15% 12 26% 13 38% 4 27%
Bankrupt 1 2% 4 12% 3 20%Total 61 54 46 34 15
Bibliography
Adams, R., and D. Ferreira, 2008, One Share-One Vote: The Empirical Evidence, Review of
Finance 12, 51–91.
Alchian, A., and H. Demsetz, 1972, Production, Information Costs and Economic Organiza-
tion, American Economic Review 62, 777–795.
Amihud, Y., 2002, Illiquidity and stock returns: cross-section and time-series effects, Journal
of Financial Markets 5, 31–56.
, and H. Mendelson, 1986, Asset Prices and the Bid-Ask Spread, Journal of Financial
Economics 17, 223–249.
Amoako-Adu, B., and B.F. Smith, 2001, Dual class firms: Capitalization, ownership structure
and recapitalization back into single class, Journal of Banking & Finance 25, 1083–1111.
Andrade, G., M. Mitchell, and E. Stafford, 2001, New Evidence and Perspectives on Mergers,
The Journal of Economic Perspectives 15, 103–120.
Ang, J.S., and W.L. Megginson, 1989, Restricted voting shares, ownership structure, and the
market value of dual-class firms, Journal of Financial Research 12, 301–318.
Arugaslan, Onur, 2007, Why are dual class shares unified?, Business Quest.
Barclay, M., and C. Holderness, 1989, Private benefits from control of public corporations,
Journal of Financial Economics 25, 371–395.
Berle, A., and G. Means, 1932, The Modern Corporation and Private Property, New York.
58
59
Berle Jr, AA, 1926, Non-Voting Stock and “Bankers’ Control”, Harvard Law Review pp.
673–693.
Bigelli, Marco, Vikas Mehrotra, and Raghavendra Rau, 2008, Expropriation, Unification and
Corporate Governance in Italy, SSRN eLibrary.
Bohmer, E., GC Sanger, and SB Varshney, 1996, The Effect of Consolidated Control on Firm
Performance: The Case of Dual-class IPOs. Empirical Issues in Raising Equity Capital
(Mario Levis, ed.), .
Boone, Audra, and J. Harold Mulherin, 2007, How firms are sold?, The Journal of Finance
62, 847–875.
Burkart, M., and S. Lee, 2008, One Share-One Vote: The Theory, Review of Finance 12,
1–49.
Chemmanur, Thomas J., and Yawen Jiao, 2007, Dual Class IPOs, Share Recapitalizations,
and Unifications: A Theoretical Analysis, SSRN eLibrary.
Claessens, S., S. Djankov, J.P.H. Fan, and L.H.P. Lang, 2002, Disentangling the Incentive
and Entrenchment Effects of Large Shareholdings, The Journal of Finance 57.
Cornett, M.M., and M.R. Vetsuypens, 1989, Voting Rights and Shareholder Wealth. The
Issuance of Limited Voting Common Stock, Managerial and Decision Economics 10, 175–
188.
DeAngelo, H., and L. DeAngelo, 1985, Managerial ownership of voting rights: A study of
public corporations with dual classes of common stock, Journal of Financial Economics
14, 33–69.
Demsetz, H., and K. Lehn, 1985, The structure of corporate ownership: causes and conse-
quences, The Journal of Political Economy 93, 1155.
60
Dewing, A.S., 1953, The financial policy of corporations (The Ronald Press).
Dimitrov, V., and P. Jain, 2006, Recapitalization of one class of common stock into dual-class:
Growth and long-run stock returns, Journal of Corporate Finance 12, 342–366.
Dittmann, I., and N. Ulbricht, 2008, Timing and Wealth Effects of German Dual Class Stock
Unifications, European Financial Management 14, 163–196.
Doidge, C., 2004, US cross-listings and the private benefits of control: evidence from dual-
class firms, Journal of Financial Economics 72, 519–553.
Ehrhardt, Olaf, Jan Kuklinski, and Eric Nowak, 2005, Unifications of Dual-Class Shares in
Germany: Empirical Evidence on the Effects of Related Changes in Ownership Structure,
Market Value and Bid-Ask Spreads, Swiss Finance Institute 06-12.
Gilson, R.J., 1987, Evaluating Dual Class Common Stock: The Relevance of Substitutes,
Virginia Law Review 73, 807–844.
Gompers, Paul A., Joy L. Ishii, and Andrew Metrick, 2008, Extreme Governance: An Analysis
of Dual-Class Companies in the United States, SSRN eLibrary.
Grossman, S., and O. Hart, 1988, One share-one vote and the market for corporate control,
Journal of Financial Economics 20, 175–202.
Harris, M., and A. Raviv, 1988, Control contests and capital structure, Journal of Financial
Economics 20, 55–86.
Hauser, S., and B. Lauterbach, 2004, The Value of Voting Rights to Majority Shareholders:
Evidence from Dual-Class Stock Unifications, Review of Financial Studies 17, 1167–1184.
Howell, Jason W., 2009, No More Share Classes: A Study of U.S. Dual Class Stock Unifca-
tions, Working Paper.
61
Jarrell, G.A., and A.B. Poulsen, 1988, Dual-Class Recapitalizations as Antitakeover Mecha-
nisms: The Recent Evidence, Journal of Financial Economics 20, 2.
Jensen, M.C., and W.H. Meckling, 1976, Theory of the Firm: Managerial Behavior, Agency
Costs and Ownership Structure, Journal of Financial Economics 3, 305–360.
Kim, Joonghyuk, Ji-Chai Lin, Ajai Singh, and Wen Yu, 2007, Dual-Class Stock Splits and
Liquidity, Working Paper.
La Porta, R., Florencio Lopez-de Silanes, and A. Shleifer, 1999, Corporate ownership around
the world, Journal of Finance pp. 471–517.
Lehn, K., J. Netter, and A. Poulsen, 1990, Consolidating corporate control: Dual-class recap-
italizations versus leveraged buyouts, Journal of Financial Economics 27, 580.
Li, Kai, Hernan Ortiz-Molina, and Xinlei Zhao, 2008, Do Shareholder Rights Affect Insti-
tutional Investment Decisions? Evidence from Dual-Class Firms, Financial Management
37.
Masulis, Ronald W., Cong Wang, and Fei Xie, 2008, Agency Problems at Dual-Class Com-
panies, Journal of Finance, Forthcoming.
Maury, Benjamin, and Anete Pajuste, 2007, Private benefits of control and dual-class share
unifications, Working Paper.
Mikkelson, W., and M. Partch, 1994, The consequences of unbundling managers’ voting
rights and equity claims, Journal of Corporate Finance 1, 175–199.
Morck, R., A. Shleifer, and R.W. Vishny, 1988, Management Ownership and Market Valua-
tion: An Empirical Analysis, Journal of Financial Economics 20, 293–315.
Nenova, T., 2003, The value of corporate voting rights and control: A cross-country analysis,
Journal of Financial Economics 68, 325–351.
62
Pagano, M., F. Panetta, and L. Zingales, 1998, Why Do Companies Go Public? An Empirical
Analysis, The Journal of Finance 53, 27–64.
Partch, M.M., 1987, The Creation of a Class of Limited Voting Common Stock and Share-
holder Wealth, Journal of Financial Economics 18, 339.
Ripley, W.Z., 1927, Main Street and Wall Street (Little, Brown, and Company).
Schleifer, A., and R.W. Vishny, 1997, A Survey of Corporate Governance, Journal of Finance
52, 737–783.
Seligman, Joel, 1986, Equal protection in shareholder voting rights: The one common share,
one vote controversy, George Washington Law Review 54, 687–725.
Smart, S.B., R.S. Thirumalai, and C.J. Zutter, 2008, What’s in a vote? The short-and long-
run impact of dual-class equity on IPO firm values, Journal of Accounting and Economics
45, 94–115.
Smart, S.B., and C.J. Zutter, 2003, Control as a motivation for underpricing: a comparison
of dual and single-class IPOs, Journal of Financial Economics 69, 85–110.
Stevens, W.H., 1926, Stockholders’ voting rights and the centralization of voting control,
The Quarterly Journal of Economics 40, 353–392.
Stulz, R., 1988, Managerial control of voting rights: Financing policies and the market for
corporate control, Journal of Financial Economics 20, 25–54.
Villalonga, B., and R. Amit, 2006, How do family ownership, control and management affect
firm value?, Journal of Financial Economics 80, 385–417.
Villalonga, Belen, and Raphael Amit, 2008, How are U.S. Family Firms Controlled?, Review
of Financial Studies, Forthcoming.
Zhang, Yi, 2002, Dual-class stock, firm value, and performance, Working Paper.
63
Zingales, L., 1994, The value of the voting right: a study of the Milan Stock Exchange
experience, Review of Financial Studies pp. 125–148.
, 1995, What determines the value of corporate votes?, The Quarterly Journal of
Economics pp. 1047–1073.
Appendix A
SEC Rule 19c-4
Rule 19c-4 – Governing Certain Listing or Authorization Determinations by National Secu-
rities Exchanges and Associations
(a) The rules of each exchange shall provide as follows: No rule, stated policy, practice, or interpre-
tation of this exchange shall permit the listing, or the continuance of the listing, of any common
stock or other equity security of a domestic issuer, if the issuer of such security issues any class
of security, or takes other corporate action, with the effect of nullifying, restricting or disparately
reducing the per share voting rights of holders of an outstanding class or classes of common stock
of such issuer registered pursuant to Section 12 of the Act.
(b) The rules of each association shall provide as follows: No rule, stated policy, practice, or interpre-
tation of this association shall permit the authorization for quotation and/or transaction reporting
through an automated inter-dealer quotation system (”authorization”), or the continuance of autho-
rization, of any common stock or other equity security of a domestic issuer, if the issuer of such
security issues any class of security, or takes other corporate action, with the effect of nullifying,
restricting, or disparately reducing the per share voting rights of holders of an outstanding class or
classes of common stock of such issuer registered pursuant to Section 12 of the Act.
(c) For the purposes of paragraphs (a) and (b) of this section, the following shall be presumed to
have the effect of nullifying, restricting, or disparately reducing the per share voting rights of an
outstanding class or classes of common stock:
1. Corporate action to impose any restriction on the voting power of shares of the common
stock of the issuer held by a beneficial or record holder based on the number of shares held
by such beneficial or record holder;
64
65
2. Corporate action to impose any restriction on the voting power of shares of the common stock
of the issuer held by a beneficial or record holder based on the length of time such shares
have been held by such beneficial or record holder;
3. Any issuance of securities through an exchange offer by the issuer for shares of an outstanding
class of the common stock of the issuer, in which the securities issued have voting rights greater
than or less than the per share voting rights of any outstanding class of the common stock
of the issuer.
4. Any issuance of securities pursuant to a stock dividend, or any other type of distribution
of stock, in which the securities issued have voting rights greater than the per share voting
rights of any outstanding class of the common stock of the issuer.
(d) For the purpose of paragraphs (a) and (b) of this section, the following, standing alone, shall
be presumed not to have the effect of nullifying, restricting, or disparately reducing the per share
voting rights of holders of an outstanding class or classes of common stock:
1. The issuance of securities pursuant to an initial registered public offering;
2. The issuance of any class of securities, through a registered public offering, with voting rights
not greater than the per share voting rights of any outstanding class of the common stock of
the issuer;
3. The issuance of any class of securities to effect a bona fide merger or acquisition, with voting
rights not greater than the per share voting rights of any outstanding class of the common
stock of the issuer.
4. Corporate action taken pursuant to state law requiring a state’s domestic corporation to
condition the voting rights of a beneficial or record holder of a specified threshold percentage of
the corporation’s voting stock on the approval of the corporation’s independent shareholders.
(e) Definitions. The following terms shall have the following meanings for purposes of this section,
and the rules of each exchange and association shall include such definitions for the purposes of the
prohibition in paragraphs (a) and (b), respectively, of this section:
66
1. The term ”Act” shall mean the Securities Exchange Act of 1934, as amended.
2. The term ”common stock” shall include any security of an issuer designated as common stock
and any security of an issuer, however designated, which, by statute or by its terms, is a
common stock (e.g., a security which entitles the holders thereof to vote generally on matters
submitted to the issuer’s security holders for a vote).
3. The term ”equity security” shall include any equity security defined as such pursuant to Rule
3a11-1 under the Act.
4. The term ”domestic issuer”shall mean an issuer that is not a ”foreign private issuer”as defined
in Rule 3b-4 under the Act .
5. The term ”security” shall include any security defined as such pursuant to Section 3(a)(10) of
the Act, but shall exclude any class of security having a preference or priority over the issuer’s
common stock as to dividends, interest payments, redemption or payments in liquidation, if
the voting rights of such securities only become effective as a result of specified events, not
relating to an acquisition of the common stock of the issuer, which reasonably can be expected
to jeopardize the issuer’s financial ability to meet its payment obligations to the holders of
that class of securities.
6. The term ”exchange” shall mean a national securities exchange, registered as such with the
Securities and Exchange Commission pursuant to Section 6 of the Act, which makes trans-
action reports available pursuant to Rule 242.601 of this chapter; and
7. The term ”association” shall mean a national securities association registered as such with
the Securities and Exchange Commission pursuant to Section 15A of the Act.
(f) An exchange or association may adopt a rule, stated policy, practice, or interpretation, subject
to the procedures specified by Section 19(b) of the Act, specifying what types of securities issuances
and other corporate actions are covered by, or excluded from, the prohibition in paragraphs (a)
and(b) of this section, respectively, if such rule, stated policy, practice, or interpretation is consistent
67
with the protection of investors and the public interest, and otherwise in furtherance of the purposes
of the Act and this section.
Appendix B
Section 313.00 of the NYSE Listed Company Manual
313.00 Voting Rights
(A) Voting Rights Policy
On May 5, 1994, the Exchange’s Board of Directors voted to modify the Exchange’s Voting Rights
Policy, which had been based on former SEC Rule 19c-4. The Policy is more flexible than Rule 19c-4.
Accordingly, the Exchange will continue to permit corporate actions or issuances by listed companies
that would have been permitted under Rule 19c-4, as well as other actions or issuances that are
not inconsistent with the new Policy. In evaluating such other actions or issuances, the Exchange
will consider, among other things, the economics of such actions or issuances and the voting rights
being granted. The Exchange’s interpretations under the Policy will be flexible, recognizing that
both the capital markets and the circumstances and needs of listed companies change over time.
The text of the Exchange’s Voting Rights Policy is as follows:
Voting rights of existing shareholders of publicly traded common stock registered under Section
12 of the Exchange Act cannot be disparately reduced or restricted through any corporate action
or issuance. Examples of such corporate action or issuance include, but are not limited to, the
adoption of time phased voting plans, the adoption of capped voting rights plans, the issuance of
super voting stock, or the issuance of stock with voting rights less than the per share voting rights
of the existing common stock through an exchange offer.
(B) Non-Voting Common Stock
The Exchange’s voting rights policy permits the listing of the voting common stock of a company
which also has outstanding a non-voting common stock as well as the listing of non-voting common
stock. However, certain safeguards must be provided to holders of a listed non-voting common
stock: (1) Any class of non-voting common stock that is listed on the Exchange must meet all
68
69
original listing standards. The rights of the holders of the non-voting common stock should, except
for voting rights, be substantially the same as those of the holders of the company’s voting common
stock. (2) Although the holders of shares of listed non-voting common stock are not entitled to vote
generally on matters submitted for shareholder action, holders of any listed non-voting common
stock must receive all communications, including proxy material, sent generally to the holders of
the voting securities of the listed company.
(C) Preferred Stock, Minimum Voting Rights Required
Preferred stock, voting as a class, should have the right to elect a minimum of two directors upon
default of the equivalent of six quarterly dividends. The right to elect directors should accrue
regardless of whether defaulted dividends occurred in consecutive periods.
The right to elect directors should remain in effect until cumulative dividends have been paid
in full or until non-cumulative dividends have been paid regularly for at least a year. The preferred
stock quorum should be low enough to ensure that the right to elect directors can be exercised as
soon as it accrues. In no event should the quorum exceed the percentage required for a quorum
of the common stock required for the election of directors. The Exchange prefers that no quorum
requirement be fixed in respect to the right of a preferred stock, voting as a class, to elect directors
when dividends are in default.
The Exchange recommends that preferred stock should have minimum voting rights even if the
preferred stock is not listed.
Increase in Authorized Amount or Creation of a Pari Passu Issue—
• An increase in the authorized amount of a class of preferred stock or the creation of a pari
passu issue should be approved by a majority of the holders of the outstanding shares of the
class or classes to be affected. The Board of Directors may increase the authorized amount
of a series or create an additional series ranking pari passu without a vote by the existing
series if shareholders authorized such action by the Board of Directors at the time the class
of preferred stock was created.
Creation of a Senior Issue—
70
• Creation of a senior equity security should require approval of at least two-thirds of the
outstanding preferred shares. The Board of Directors may create a senior series without a
vote by the existing series if shareholders authorized such action by the Board of Directors
at the time of the existing series of preferred stock was created.
• A vote by an existing class of preferred stock is not required for the creation of a senior issue
if the existing class has previously received adequate notice of redemption to occur within 90
days. However, the vote of the existing class should not be denied if all or part of the existing
issue is being retired with proceeds from the sale of the new stock.
Alteration of Existing Provisions—
• Approval by the holders of at least two-thirds of the outstanding shares of a preferred stock
should be required for adoption of any charter or by-law amendment that would materially
affect existing terms of the preferred stock.
• If all series of a class of preferred stock are not equally affected by the proposed changes,
there should be a two-thirds approval of the class and a two-thirds approval of the series that
will have a diminished status.
• The charter should not hinder the shareholders’ right to alter the terms of a preferred stock
by limiting modification to specific items, e.g., interest rate, redemption price.
SUPPLEMENTARY MATERIAL
.10 Companies with Dual Class Structures —
The restriction against the issuance of super voting stock is primarily intended to apply to the
issuance of a new class of stock, and companies with existing dual class capital structures would
generally be permitted to issue additional shares of the existing super voting stock without conflict
with this Policy.
.20 Consultation with the Exchange —
Violation of the Exchange’s Voting Rights Policy could result in the loss of an Issuer’s Exchange
71
market or public trading market. The Policy can apply to a variety of corporate actions and securi-
ties issuances, not just super voting or so-called ”time phase”voting common stock. While the Policy
will continue to permit actions previously permitted under Rule 19c-4, it is extremely important
that listed companies communicate their intentions to their Exchange representatives as early as
possible before taking any action or committing to take any action that may be inconsistent with
the Policy. The Exchange urges listed companies not to assume, without first discussing the matter
with the Exchange staff, that a particular issuance of common or preferred stock or the taking
of some other corporate action will necessarily be consistent with the Policy. It is suggested that
copies of preliminary proxy or other material concerning matters subject to the Policy be furnished
to the Exchange for review prior to formal filing.
.30 Review of Past Voting Rights Activities —
In reviewing an application for initial listing on the Exchange, the Exchange will review the issuer’s
past corporate actions to determine whether another self-regulatory organization (”SRO”) has found
any of the issuer’s actions to have been a violation or evasion of the SRO’s voting rights policy.
Based on such review, the Exchange may take any appropriate action, including the denial of the
listing or the placing of restrictions on such listing. The Exchange will also review whether an issuer
seeking initial listing on the Exchange has requested a ruling or interpretation from another SRO
regarding the application of that SRO’s voting rights policy with respect to a proposed transaction.
If so, the Exchange will consider that fact in determining its response to any ruling or interpretation
that the issuer may request on the same or similar transaction.
.40 Non-U.S. Companies —
The Exchange will accept any action or issuance relating to the voting rights structure of a non-U.S.
company that is in compliance with the Exchange’s requirements for domestic companies or that
is not prohibited by the company’s home country law.
Appendix C
Sample of Dual Class Firms
In order to find a comprehensive sample of dual class firms, I first pull all firm years from
the CRSP/Compustat Merged Fundamental Annuals database for fiscal years 1988 to 2007.
I apply four criteria to the sample: 1) revenues or assets of at least one dollar, 2) sic code
is not equal to 6798, 3) the firm is American, and 4) the firm’s stock trades on the NYSE,
AMEX, NASDAQ, or over-the-counter (Exchange (EXCHG) codes 11, 12, 13, 14, and 19).
This results in a sample of 13,971 firms and 113,866 firm years.
Using the full sample of 113,866 firm years, I test each firm year against a series of five
dual class tests:
Compustat class test - Firms with “CL A” or “CL B” in the company name field pass
the Compustat class test.
Center for Research Prices class code test (CRSP code test) - A firm which has
a share class code in the shrcd field passes the CRSP code test.
Securities Data Corporation test (SDC test) - I pull all multiple class issues (both
IPO and SEO - a total of 1,444) from SDC (01/01/1980 to 12/31/2007). A firm passes the
SDC test if the fiscal year end date is after the issue date listed in SDC.
IPO test- I use the list of initial public offerings by dual-class firms from Jay Ritter’s
IPO website (Smart and Zutter collected most of the data). The list contains 570 firms who
went public between (01/01/1980 to 12/31/2007). A firm passes the IPO test if the fiscal
year end date is after the issue date identified by Ritter.
CRSP/Compustat shares outstanding test - CRSP and Compustat both maintain
records on the number of shares outstanding; however, CRSP’s shares outstanding figure is
72
73
based on a specific share class, whereas Compustat’s shares outstanding figure is based on a
total of all shares outstanding (all classes included). By comparing the shares outstanding,
dual class firms may be identified. This method was first used by Zhang(2008) and has
also been used by Gompers, Ishii, and Metrick(2008). A firm passes the CRSP/Compustat
shares outstanding test if it meets a 1% threshold and certain frequency criteria (to eliminate
random errors).
If the firm passes any one of the test in any firm year, I classify the firms as potentially
dual-class. Once the firm is identified as a potential dual class firm, all of its firm years are
included in the test sample. The resulting test sample is made up of 3,410 firms and 27,738
firm years.
From this test sample, I examine annual filings with the Securities and Exchange Com-
mission to determine if and when the firm had a dual-class stock structure. The final sample
consists of 1,096 firms which use the dual class structure and 8,245 firm years.
Appendix D
Sample Unification
This appendix presents information regarding a dual class unification at E-Z-EM (AMEX:EZM).
Information comes from news articles on Factiva or Lexis-Nexis and firm proxies.
Unification Timeline
Dual class recapitalization announcement September 29, 1992
Class B (EZM.B - non-voting) shares begin trading October 27, 1992
Board begins to examine unification options October 2001
A committee of outside directors begin to evaluate unification May 6, 2002
Board recommends unification July 9, 2002
Announcement of the proposed unification by press release July 10, 2002
Proxy statement mailed discussing unification September 13, 2002
Unification approved by shareholders October 15, 2002
New common stock share begins trading October 22, 2002
Dual class structure details:
Class A Common Stock Terms:
• Voting: One vote per share. 66% affirmative vote of Class A shares actually voted required
for any amendment of the certificate of incorporation, reduction of capital, merger with and
into one or more corporations, sale, transfer, pledge, etc. of substantially all of the Company’s
property or assets, or liquidation, dissolution or winding up of the Company.
• Dividends: May receive cash dividends equal to or less than dividends paid on Class B common
stock. May receive stock dividends either in the form of Class A or Class B common stock.
Class B Common Stock Terms:
74
75
• Voting: No vote.
• Dividends: May receive cash dividends equal to or greater than dividends paid on Class A
common stock. May receive stock dividends only in the form of Class B common stock.
• Conversion: May be converted into Class A common stock on a one-for-one basis if either
– the Class A or Class B shares are excluded from quotation on the AMEX due to the
dual class structure, or
– the number of outstanding shares of Class A common stock falls below 10% of total
number of shares of all classes of outstanding E-Z-EM common stock.
Why was the structure implemented?
In their 2002 annual proxy, E-Z-EM gives the following reasons why the dual class structure was
originally implemented:
• to allow E-Z-EM to issue equity securities in connection with acquisitions and to raise equity
capital or to issue convertible debt or convertible preferred stock as a means to finance future
growth without diluting the voting power of the Company’s existing stockholders;
• to allow E-Z-EM to grant equity-based compensation awards without diluting the voting
power of the Company’s existing stockholders;
• to allow the existing holders of E-Z-EM common shares to sell or otherwise dispose of common
shares while maintaining their voting positions; and
• to reduce the risk of an unsolicited takeover attempt that might not be in the best interests
of the Company and its stockholders.
76
Why was the structure being discarded?
Also in their 2002 annual proxy, E-Z-EM states the elimination of the structure is expected to:
• eliminate potential investor confusion and additional administrative expenses caused by our
dual class capital structure,
• eliminate any negative impact on the market price of shares that we believe results from the
dual class structure,
• potentially increase our investor base and the liquidity, trading volume and trading efficiencies
of our common shares,
• potentially increase our ability to use stock as an acquisition currency, and
• potentially enhance our ability to attract analyst coverage and investments by mutual funds
and other types of investors that do not purchase non-voting securities.
Shares outstanding and control block:
As of the record date for their 2002 annual meeting, E-Z-EM had approximately 4,001,341 shares of
(voting) Class A common stock outstanding, and 5,990,974 shares of (non-voting) Class B common
stock outstanding. The Stern and Meyers families held approximately 64% of the voting Class A
stock and 52% of the non-voting Class B common stock.
Who can vote for the unification proposal?
Only Class A shareholders have the right to vote for the unification proposal.1
Unification Details:
In E-Z-EM’s unification, each share of Class A common stock and each share of Class B common
stock was converted into one share of new common stock. In E-Z-EM’s case, the company actually
did a recapitalization merger with a wholly-owned subsidiary to effect the unification.
1Sometimes both classes are allowed to vote for the proposals.